California Management Review

Digital transformation has given rise to open models of innovation, distributed forms of organization, and shared business models. Brands, however, have remained confined to the traditional boundaries of their firms. Considered by some as a firm’s most valuable resource, brands attract investors and talent, provide heuristics for the quality of goods and services provided to consumers, and foster loyalty. Brands are built around a name, design, color, symbol, or any artwork that serves as a cue of corporate culture, product quality, or sometimes personality. As a source of competitive advantage, firms protect these aesthetic features of their brands with trademarks and copyrights to build and maintain a brand reputation.1 This prevents third parties from using these brands in contexts other than those orchestrated by the firm or for other products and services. As they grow, firms scale their brands as ‘houses of brands’ or ‘branded houses.’1 (Figure 1a and 1b) Figure 1. a) Branded House of Virgin, b) House of Brands of P&G, and c) Web of Brands of Nouns ‘House of brands’ architecture is organized around a core corporate brand with many subbrands. For instance, Procter & Gamble owns 65 distinguishable subbrands, such as Pampers, Gillette, and Tampax. With this model, firms can experiment with new products and markets while protecting their core brands from reputational damage. It can be difficult, however, to instill a new brand into consumers’ minds. In ‘branded house’ architecture, a core brand is directly associated with subbrands. Virgin Media, Virgin Galactic, and Virgin Mobile, for instance, all share the Virgin name. This direct association facilitates market entry by leveraging the reputation of the core brand but discourages experimentation, as any faux pas can hurt the core brand. With both architectures, firms manage their house brands behind closed doors, taking a top-down approach to decision-making that trickles to their subbrands. Managers’ are tasked with designing an appropriate organizational structure to deploy their brand portfolios and attracting the right talent and skills to execute their decisions. In other words, the strategic management of brands is mostly performed behind closed doors within the house. We believe that emerging technologies offer new avenues to build, manage and grow brands, and we call such brands ‘open brands’. Using blockchain technology and governed by online communities, open brands embrace open-source principles and harness the wisdom of the crowd to grow. This model turns the process of brand building upside down, akin to models of innovation that have moved from closed, top-down processes to open, bottom-up processes. First, we flesh out the main characteristics of an open brand. Second, we illustrate this model with Nouns—an experiment built on the Ethereum network, a blockchain technology. We show how Nouns provides a model for managing brands with online communities at the core. Finally, we highlight opportunities for open brands and the challenges faced in leveraging them. Open brands The ‘open brand’ concept sounds like an oxymoron. Corporate brands are usually considered ‘closed’, that is, carefully built and protected within a firm’s boundaries. Instead of being characterized by firms, hierarchical governance, and legal protections, open brands are characterized by communities, decentralized decision-making, public-domain artwork, and the use of blockchain technology. Community-owned brands and webs of brands Whether it is organized as a branded house or a house of brands, firms traditionally own house brands. In contrast, open brands blur the barrier between firms and consumers by shifting ownership to online communities. Open brands are initially created by a small group of founders and managed by a progressively larger crowd. Members acquire affiliation with the brand—often in the form of a digital asset—and bring their knowledge and social capital, turning the exercise of brand-building into a form of collective action. As such, consumers can acquire decision rights over the common treasury and contribute to the scaling activities of the open brand. Instead of an expanding house, open brands develop into ‘webs of brands’ (Figure 1c). Community members self-organize around a set of common assets (brand name, design, treasuries, etc.) to determine the direction of the brand.3 Protocols facilitate the coordination of the community, and any member can launch new product lines, suggest growth strategies or create subopen brands with their own communities. Each project extends the open brand in a network-like fashion by rallying a faction of the community and tapping into new audiences. Ultimately, open brands crowdsource products and strategies thereby benefit from the experimentation afforded by the house of brands model but maintain their identity as branded houses thanks to public -domain artwork and blockchain technology. Public-domain artwork and blockchain Open brands have artwork, fonts, and various distinguishable aesthetic features in the public domain. Anyone can copy, modify, and recombine them into marketable products, services or subopen brands. Effectively, public-domain artwork functions similar to open-source code or open-source hardware, providing a platform on which to experiment and innovate.4 The more relatable and remixable art is, the more likely users will be able to identify with the open brand and reuse it in various contexts. In other words, open brands use public-domain artwork with a high level of “meme-ability” potential. While aesthetic features are available to anyone for use, open brands also have ways to build reputations and legitimacy thanks to blockchain technology. Indeed, votes cast by the community are recorded on a public blockchain, allowing any user to verify which projects have received the approval of the open-brand community. Furthermore, blockchain technology is used to track the provenance of digital assets, allowing users to distinguish official uses from counterfeited uses. As such, open brands rely on technology instead of legal instruments to protect their reputations and build a legacy. Open brands are vested with powerful scaling and incentive mechanisms that traditional brands do not enjoy. First, open brands are more similar to memes –humorous images or videos modified and sent via the internet – than to traditional brands insofar as they actively invite experimentation with their aesthetic features instead of preventing it. As such, any iteration of this artwork increases awareness of the original one in the same fashion that academic citations give credit to and build upon foundational papers.5,6 Second, the community members of open brands are both intrinsically and extrinsically motivated to scale. Indeed, membership not only facilitates fun and social belonging but also allows owners and community members to benefit financially from the success of the open brand. This diminishes the classical agency costs present in principal-agent situations with house of brands or branded house models. Finally, and relatedly, the radical transparency and inclusiveness of open brands brings legitimacy to the direction of the brand and helps form a sense of common identity within the community that fuels participation, loyalty, and authenticity.7 The case of Nouns Nouns is an open brand with serious ambitions. The defining aesthetics of the brand are cartoon characters with square-framed glasses—Nouns—that can be used as avatars in online communication channels and social media (see Figure 2). The project was launched by a team of 10 contributors, some of whom are anonymous, such as 4156, and some of whom are known individuals, such as Dominik Hofmann, a cofounder of Vine. Figure 2. Eight different Nouns randomly generated and auctioned by an algorithm Nouns have appeared on a Budweiser Super Bowl advertisement and in the contexts of e-sport teams and products such as streetwear, skateboards, and comics. The success of Nouns lies in their highly memeable aesthetics in the public domain, their use of blockchain protocols, and their decentralized governance mechanisms. CC0 artwork Each Noun is the product of a generative algorithm that creates ‘one Noun every day, forever’. The components (glasses, background, body, accessories, and head) of Nouns were designed by a group of artists including Gremplin and the collective eBoy Arts, among others. The algorithms can build millions of unique Noun combinations, and the experiment can run for hundreds of thousands of years. Importantly, the art used to design Nouns is under CC0 licensing, that is, ‘no rights reserved’ licensing, meaning that anyone can copy, modify, and commercially reuse the art. As such, the Noun glasses quickly became the defining aesthetic of the open brand and have been reused in a great variety of contexts (Figure 3). Hundreds of derivative (subbrand) projects have adopted a “Nounish” look by using square-framed glasses and transposing this concept to other physical and digital goods.8 The combination of CC0 licensing and remixable and distinguishable artwork is the first set of ingredients that explains the success of Nouns. Figure 3. Noun glasses making their appearance in the Budweiser Superbowl advertisement Nouns DAO Anyone can right-click and save a picture of a Noun, but instead, people pay 5 figures to acquire an original one, certified with blockchain technology, through an auction mechanism. All the proceeds go to a treasury managed by a decentralized autonomous organization (DAO)—the Nouns DAO. People are willing to pay for original Nouns because the digital asset grants a vote in the management of the treasury. Additionally, if their open brands succeed, Noun owners expect the value of their Nouns to increase. The Nouns DAO is the governing body that can change the art-generating protocol and manage the treasury resulting from sales of Nouns. At its core, the collective is an online community made of anonymous members organized as a DAO. DAOs resemble institutions in terms of collective action because they manage a common pool of resources;9 however, they differ in their use of decentralized and trustless governance mechanisms. Specifically, these mechanisms rely on smart contracts—automated algorithms executed on the Ethereum blockchain—to manage memberships and voting. In its existence of slightly over a year, the Nouns DAO has amassed a treasury worth more than 24k ETH (approximately 40 million USD) and a core community of more than 290 members. More than 150 proposals have been made, and 117 have been executed, including donations to charities, the development of product lines such as clothing or comics using the Noun brand, and the financing of NounDAO operations. Additionally, retroactive funding has been granted for projects and individuals who have built successful initiatives for the growth of the ecosystem. For instance, 1000 ETH (approximately 1.2 million USD) has been allocated to NounsBuilder, a solution allowing anyone to replicate the Nouns model (art generation, auction, governance), thereby inviting anyone to experiment with open brands. Nouns’ year-long experiment offers some valuable knowledge regarding open brands. First, the artwork started as profile pictures with which individuals or companies could build online identities. However, the square-framed glasses quickly proved to be a versatile resource that could be deployed in contexts as varied as software development, gaming, merchandising, or beverages. Relatable and remixable artwork in the public domain attracts large communities and invites reuse. Second, Nouns leverages digital assets on a blockchain to incorporate incentives for community participants. Members not only join the DAO for experimentation and fun but can also be financially rewarded for participation in the growth of the brands. Finally, 4156, a co-founder of Nouns, highlighted the absence of contracts signed between the Nouns Foundation and Bud Light10 in their collaboration. In some cases, blockchains could substitute for formal modes of governance such as employment or partnership agreements. Open brands can also be a tool to build legitimacy and reputation. Plug-or-play and challenges Open brands are here to stay, and traditional firms have two avenues to experiment with them: plug or play. The plug strategy consists of becoming a member of an open-brand community. This approach allows a firm to connect with – or to plug into - an existing audience and leverage its dedication to guerrilla marketing. It is the strategy embraced by Budweiser in its promotion of Bud Light Next – its first zero-carbohydrate beer promoted in collaboration with the NounDAO. By doing this, Budweiser displayed support for the Ethereum community and became an active member of the NounDAO, participating in 75 different proposals. The play strategy involves the creation of an open brand and experimentation with a firm’s existing audience. Firms subsidize the ingredients to bootstrap the open brand (aesthetics, governance rules, discussion platform, treasury) in the same fashion that they would provide toolkits to foster user innovation.11 They then play the role of community managers - conveying information and providing support - and let members play with their open-brand resources. Open brands not only carry powerful promises but also come with a set of trade-offs. First, permissive licenses accompanying the artwork of open brands facilitate fast, meme-like diffusion but bring challenges in the management of brand reputation and identity. Because this artwork can be remixed by anyone, audiences may find it challenging to delineate a single culture and set of values embodied by an open brand. This leads to openness vis-à-vis the interpretation of the open brand and associated subbrands, which can prove to be a double-edged sword.12 Fortunately, the use of blockchain technology as a transparent record allows us to trace back original artwork, but it could be a barrier for less technically inclined individuals. As such, brand recognition is no longer solely guaranteed by protected aesthetics features but instead protected through the use of immutable digital records. Second, open brands are managed by online communities known for their fluidity in membership. These allow DAOs to leverage both the knowledge and social capital of members13 for the construction and growth of brands; however, sustaining continuous and strong member engagement can be challenging. Nevertheless, the use of digital assets as membership tickets to join DAOs provides powerful intrinsic and extrinsic motivation bridges to sustain participation in the community. Furthermore, the permissionless nature of online communities potentially brings a wide variety of ideas at the expense of coordinating costs and reduced speed.14 Blockchain technology such as Ethereum render these coordinating costs tangible via the use of ‘gas fees’—up to a few hundred dollars per transaction—but scaling solutions offer promising avenues to reduce them. Third, it remains unclear how transparent discussions on the positioning and direction of open brands—accessible by anyone—affects their competitive positioning. For instance, in the context of open innovation, it has been shown that openness benefits organizations at an early stage of development but that these organizations typically revert to more closed forms as they mature.15 In the case of open brands, it could very well be that the secret to their success is not to be found in the content of their strategies but in the competencies of the people joining the management of the commons.16 Acknowledgments I would like to thank Prof. Scott Duke Kominers (Harvard Business School) and Prof. Patricia Wolf (University of Southern Denmark) for their helpful guidance during the writing of this article. Also, I would like to thank Andrea Lenzner and Dr. Julian Müller for their careful review and comments to improve earlier versions of this manuscript. All remaining errors are mine. Funding This project did not receive any funding. Conflicts of interest The author owns Ethereum cryptocurrency. References J. Barney, “Firm Resources and Sustained Competitive Advantage,” Journal of Management 17, no. 1 (March 1991): 99-120. D.A. Aaker and E. Joachimsthaler, “The Brand Relationship Spectrum: The Key to the Brand Architecture Challenge” California Management Review 42, no. 4 (July 2000): 8-23. Ø.D. Fjeldstad, C.C. Snow, R.E. Miles, and C. Lettl, “The Architecture of Collaboration,” Strategic Management Journal 33, no. 6 (June 2012): 734-750. E. Von Hippel, “User Toolkits for Innovation,” Journal of Product Innovation Management 18, no. 4 (July 2001): 247-257. https://a16zcrypto.com/cc0-nft-creative-commons-zero-license-rights/ https://decrypt.co/106761/why-ethereum-nft-creators-are-giving-away-commercial-rights-to-everyone J. Hautz, D. Seidl, and R. Whittington, “Open Strategy: Dimensions, Dilemmas, Dynamics,” Long Range Planning 50, no. 3 (June 2017): 298-309. https://nouns.center/projects E. Ostrom, “Governing the Commons: The Evolution of Institutions for Collective Action,” (Cambridge: Cambridge University Press, 1990). https://decrypt.co/90860/bud-light-nouns-ethereum-nft-superbowl-ad E. Von Hippel and R. Katz, “Shifting Innovation to Users via Toolkits,” Management Science 48, no. 7 (July 2002): 821-833. T. Meyvis and C. Janiszewski, “When are Broader Brands Stronger Brands? An Accessibility Perspective on the Success of Brand Extensions,” Journal of Consumer Research 31, no. 2 (September 2004): 346-357; S.J. Milberg, C.W. Park, and M.S. McCarthy, “Managing Negative Feedback Effects Associated with Brand Extensions: The Impact of Alternative Branding Strategies,” Journal of Consumer Psychology 6, no. 2 (January 1997): 119-140. H. Safadi, S.L. Johnson, and S. Faraj, “Who Contributes Knowledge? Core-Periphery Tension in Online Innovation Communities,” Organization Science 32, no. 3 (May 2020): 752-775. D.P. Ashmos, D. Duchon, R.R. McDaniel, and J.W. Huonker, “What a Mess! Participation as a Simple Managerial Rule to ‘Complexify’ Organizations,” Journal of Management Studies 39, no. 2 (March 2002): 189-206. M.M. Appleyard and H.W. Chesbrough, “The Dynamics of Open Strategy: From Adoption to Reversion,” Long Range Planning 50, no. 3 (June 2017): 310-321. N.J. Foss, P.G. Klein, L.B. Lien, T. Zellweger, and T. Zenger, “Ownership Competence,” Strategic Management Journal 42, no. 2 (February 2021): 302-328.

COVID-19 has dramatically accelerated the process of digitalization in developed economies, bringing a multitude of benefits to firms and workers alike. Career guidance and recruitment services are amongst those with great potential to benefit from these dynamics. However, few businesses in this area have capitalized fully on the opportunities offered by digitalization and, in particular, artificial intelligence. We argue that a hybrid model which combines digital and in-presence career support is displaying advantages in terms of efficiency and effectiveness of the services provided, and should be embraced more readily by firms offering recruitment and career guidance services to job seekers. This view is in line with the supposition that we are at the beginning of a transition into the so-called ‘feeling economy’, expected to become dominant within the coming two decades, whereby the feeling tasks of jobs will become the core responsibility of human workers, while the thinking task of jobs will be relegated to artificial intelligence (AI) (Huang et al., 2019). Recruitment and career service providers can engage with these changes by being more people oriented and honing their soft skills, while leaving the increasingly-complex data management tasks to AI. Furthermore, in the context of the post-COVID-19 economic environment, and the entry of Gen-Z to the workplace, there is ample scope for positive contribution from avant-garde career guidance and recruitment organization to a bruised and arguably fragile labor force. When guidance centers remained closed for months or operated at reduced service levels due to COVID-19, job candidates were left with insufficient support. COVID-19 has also negatively affected industries and worsened the economic recovery of young and unemployed people, generating large numbers of individuals needing career guidance and recruitment support. New and more complex challenges faced by job seekers, and by job recruiters and employers, cannot be addressed solely using the tools and approaches of the past, but should be tackled in a faster and a more personalized way through the newest digital and AI tools available. Precisely this is the primary goal of AI, which is usually described as a system with the “ability to interpret external data correctly, to learn from such data, and to use those learnings to achieve specific goals and tasks through flexible adaptation” (Haenlein and Kaplan, 2019). Within the Information, advice, and guidance (IAG) industry, and the greater ecosystem which includes recruitment firms and HR departments, the technical infrastructure to provide guidance and the recruitment services with the assistance of emerging technologies has been established to some degree, but what such service providers still lack is a more strategic vision on digitalization and integration of AI solutions, which promise to benefit all stakeholders in the industry. In this article, we first look at the post-COVID-19 labor market dynamics and a number of defining characteristics of the current labor force. Subsequently, we outline the key attributes of digital and AI solutions in the career guidance and recruitment service business model, drawing attention to the benefit that such solutions promise for both, service providers and job seekers. Post-COVID-19 labor market dynamics   Figure 1 shows how the unemployment rate in the United States evolved from January 1994 to March 2023 differentiating between categories of un- or under-employed workers in addition to the standard measure of unemployment. Specifically, these include discouraged workers, e.g., individuals “not currently looking for work specifically because they believed no jobs were available for them or there were none for which they would qualify” (U.S. Bureau of Labor Statistics, 2023) or persons marginally attached to the labor force, e.g., “persons not in the labor force who want and are available for work, and who have looked for a job sometime in the prior 12 months” (ibid.), or involuntarily part-time employed workers “because of an economic reason, such as their hours were cut back or they were unable to find full-time jobs” (ibid.). In this specific regard, the unemployment rate has jumped dramatically during the first wave of the COVID-19 pandemic (spring/summer 2020) exceeding the peak reached in the acute phase of the global economic and financial crisis (end of 2009/beginning of 2010). Furthermore, Figure 1 supports the view that the phenomenon of labor force marginalization, under-employment or abandonment of the labor force by discouraged workers is cyclical and (increasingly) extreme, impacting heavily on the labor market as well as recruitment and career guidance services.  Figure 1: Unemployment rate after considering discouraged workers, marginally attached persons and employed part time for economic reasons), January 1994 - March 2023 (Source: own elaboration based on Federal Reserve Bank of St. Louis (2023)) This data pattern can be further elaborated on by considering two well-established labor market themes, which in turn highlight how AI-driven career guidance can provide support for both, the service providers as well as for job seekers: the Great Resignation: COVID-19 greatly increased the number of voluntary resignations as employees re-assessed their career priorities and choices (Groysberg et al., 2018). It is beyond scope to investigate in detail the underlying reasons for this trend, which may include disillusionment/discouragement, dissatisfaction with prior employment or a post-pandemic shift of values toward a better work-life balance, or a mutation of the so-called psychological contract, i.e., what employees expect of their employers, and vice versa (Schroth, 2019). However, Table 1 shows that individuals have increasingly (voluntarily) quit their jobs from the end of 2020 onwards. The Great Resignation poses an important challenge for the career guidance sector, which is now dealing with increasing volumes of job candidates with a great variety of expectations as well as professional and psychological profiles, for numerous job seekers on a quest for well-being and in search of new and better job opportunities, and for recruiters having to evaluate a glut of applications and to attract the most suitable applicants. As we will see, career guidance and recruitment services can capitalize on AI to improve effectiveness and efficiency of the service they provide, benefitting their own firms and candidates alike; entry into the labor market of Generation Z: individuals born from 1995 to 2010 have grown up in an increasingly interconnected, digitalized world and expect seamless and technology-driven solutions in various spheres of their lives, including career guidance and job search (Twenge, 2017). Simultaneously, they are entering the job market with far less work experience than previous generations, possibly as a result of growing up in more affluent households or perhaps because they are forced to spend summers on extracurricular and preparatory activities, given the tougher competition for entry into institutions of higher education (Schroth, 2019). Furthermore, statistics paint a far-from-optimistic picture of the mental health and resilience of the Gen Z generation, which suffers from the highest rate of depression and anxiety, in spite, or perhaps precisely because of, having grown up in a ‘culture of safety’ and parental overprotection. At the same time, this psychosocial fragility translates into an increased sensitivity and passion for issues of social justice, making Gen Z the generation most concerned with diversity, equity and inclusion (ibid.). AI-driven career guidance and recruitment platforms can assist in catering to the nuance and demanding expectations of this generation by offering personalized, more efficient and data-driven support, while leaving human career and recruitment agents the time and energy to support candidates through meaningful one-one-one interactions. Undoubtedly, job insecurity combined with the pressure to find fulfilling job opportunities can have a significantly negative impact on mental health (Paul and Moser, 2009). AI-driven career guidance can help to reduce stress by better streamlining the job search process and by assisting in the identification of employment opportunities aligned with evolving personal interests and values (Pentina and Tarafdar, 2014) of a sensitive, fragile, socio-conscious and complex generation of workers. In such a multifaceted and rapidly evolving environment, AI-driven career guidance and recruitment has the potential to address key challenges involving employment volatility, the Great Resignation and unique needs of Generation Z. While forward-looking companies in the career support sector are committed to investing in innovation, the true challenge is reshaping the traditional career guidance and recruitment model based on low levels of cutting-edge technology into a new paradigm where sophisticated technology and individuals can effectively meet. This is what authors like Barnes et al. (2010) and Bårdsdatter Bakke et al. (2018) pioneeringly refer to as “e-guidance”. Table 1: total nonfarm quits (2001-2022) and the Great Resignation after COVID-19 (Source: own elaboration based on Federal Reserve Bank of St. Louis (2023))  This new approach envisages the provision of a more streamlined, comprehensive and cutting-edge career guidance and recruitment service in the absence of physical space constraints and traditional business hours. Personalizing these services (i.e., providing different services based on the varying needs of job candidates) and thus dramatically expanding the user base is crucial to sustaining the economic validity of the necessary investments in the underlying technological tools, use of which by definition expands the number of job candidates that can be supported and matched with suitable jobs (and vice versa) simultaneously. Key attributes and benefits of the e-guidance model As shown in Figure 2, the AI-driven career guidance and recruitment model has six core features that are key in the provision of added value to all stakeholders in the IAG and recruitment industry: scalability: AI-based career guidance and recruitment can provide personalized advice and serve numerous job seekers simultaneously, while ensuring that they have access to any support needed through one-on-one meetings with counselors. AI is rapidly improving in its capacity to automate data-based ‘thinking’, or analytical, tasks, leaving room for human workers to engage in meaningful ‘feeling’ activities that require emotional intelligence and the interactive nuance in which human beings have an obvious upper hand (Huang et al., 2019), at least for the foreseeable future. Such scalability promises one step in the right direction to the ultimate aim of making IAG and recruitment services a right rather than a privilege, and ensuring that they are available to all individuals who require them, in line with the ambitious objectives set out by the International Labor Organization (ILO 2021);  real-time job market analysis: the current trajectory in AI capability development furthermore implies that real-time collection and analysis of large amounts of job market data will become ever-more accessible to career guidance and recruitment firms, allowing them to support job seekers more effectively by identify demanded skills and emerging industries, and adapting their advice accordingly; efficient job matching and personalization: in addition to identifying in-demand skills and emerging industries, AI-driven career guidance platforms can rapidly identify relevant job opportunities based on a job seeker’s detailed profile, suggesting that the job search process can be dramatically sped up and personalized, and chances of a candidate finding a suitable position improved. AI’s absolute advantage over human counterparts in handling large amounts of data suggests that candidate profiles can be made more elaborate by taking into account their skills, interests, values and personal circumstances, without necessarily compromising a candidate’s chance of finding suitable employment. Smart technologies can provide additional and up-to-date market intelligence information to support job seekers in identifying the most suitable source of employment in terms of location, company, job requirements (i.e., skills etc.) or benefits (i.e., wage, contract typology etc.). Since collecting, categorizing, filling out and re-elaborating information on vacancies and training opportunities from multiple sources is a relatively simple task for a well-instructed machine, AI-technology allows career guidance practitioners to focus on added-value in providing support to job seekers through mentoring, motivating and otherwise advising job candidates. As the Feeling Economy (Huang et al., 2019) becomes more obviously dominant in the future, human agents in the guidance and recruitment process will be left with more resources to focus on building constructive and supportive intra-personal relationships with candidates. AI’s major impact thus consists of complementing and increasing the human abilities of guidance practitioners. Digitalization can guarantee a more sustainable practitioners-to-job-seekers ratio without hampering service quality, squeezing private sector margins or increasing public expenditure. enhanced self-assessment: on the basis of powerful AI analytical capacity, e-guidance platforms can provide more comprehensive and objective self-assessment tools for candidates, allowing job seekers to gain a better understanding of their strengths and weaknesses, and to undertake self-analysis in the spirit of the SWOT approach (strengths, weaknesses, opportunities and threats) frequently utilized in business processes (Agrawal et al., 2018); bias reduction: the importance of reducing bias in the job market and of incorporating the less advantaged labor market participants into mainstream jobs has been a focal discussion in management literature for decades; see, for instance, the seminal work of J. L. Koch, published in the California Management Review in 1974. The topic is likely to gain increasing attention in future as the new generation enters the workforce. As mentioned previously, for Gen Z issues of diversity, equity and inclusion are more conspicuous than for any previous generation (Schroth, 2019). While the reiteration of biases is a well-established concern surrounding the topic of artificial intelligence in decision making, it is worth noting that carefully-designed AI can actually contribute to the reduction of such biases, as for example in the career guidance and recruitment services. AI-based career guidance can potentially reduce human biases (Moore, 2017) characterizing “traditional” career counseling and lead to job search outcomes perceived by job seekers from heterogeneous backgrounds as more objective. As effectively summed up by Windley (2022), “AI doesn’t get tired or accidentally ignore a qualified candidate. On the contrary, every candidate is given an equal and fair look, regardless of the time of day, the stress level of the hiring manager or any number of human conditions that can create variants when reviewing résumés”. AI-based career guidance can be a powerful tool in reducing the digital divide while fostering the inclusion of minorities, neurodiverse individuals, or women facing societal barriers; continuous learning and improvement: AI agents can learn from social interactions, increasing their understanding of a given context and the people within it, and consequently producing more meaningful and suitable output (Krishna et al., 2022). This capability can be capitalized on in the context of career guidance and recruitment, whereby AI-based platforms learn from user interactions and continuously improve their career and job recommendations over time. Figure 2: Main opportunities of AI-driven career guidance for job seekers. In sum, job seekers need interfaces reflecting their daily use of online technologies and would benefit from an integration of AI in career guidance and recruitment platforms. Such platforms should offer intuitive and straightforward user interfaces to ensure that job seekers do not suffer unnecessary technology-induced stress that may negatively impact on their motivation levels during the inherently challenging job search process. If wisely implemented, AI-driven e-guidance presents the potential to provide personalized, scalable, data-driven, continuously improving support to job seekers, while reducing potential biases based on race/ethnicity, gender and sexual orientation, age, disability, religion, neurodiversity and economic status. Final considerations From the perspective of job seekers, one of the major challenges of ongoing digitalization is the phenomenon of the “digital divide”. An increasing degree of digitalization in the IAG and recruitment industry can be highly beneficial for people with basic digital skills and access to the necessary infrastructure, because it enables faster connectivity to more efficient and effective services. For instance, individuals suffering from some form of isolation (e.g., geographical, health-related etc.) or facing significant time constraints (e.g., as a result of burdensome care responsibilities toward relatives), can benefit by avoiding lengthy commutes. Digital connectivity and skills are interconnected with socioeconomic status and, therefore, it is essential to carefully evaluate the pros and cons of digitalization for different target groups. Moreover, it seems necessary to design differentiated strategies of interaction with job seekers, which might best cater to increasingly heterogeneous needs. Introducing digital career-guidance and recruitment tools in parallel to in-person services represents the solution to reach new levels of efficiency, capacity and effectiveness of the services provided. However, excessive reliance on digital services at the expense of in-person relationships represents a significant, not-to-be-neglected risk, especially for individuals without digital infrastructure and/or skills, who are also amongst the most vulnerable job seekers needing support in re-entering the rapidly changing and increasingly complex job market. This is precisely why AI-driven platforms, by increasing scale, efficiency and equity, and by enriching the information content, and thus upscaling personalization, of recruitment and career guidance advice, can make a positive difference. References Agrawal, A., Gans, J. S., & Goldfarb, A. (2018). Prediction machines: the simple economics of artificial intelligence. Harvard Business Press. Bårdsdatter Bakke, I., Hagaseth Haug, E., & Hooley, T. (2018). Moving from information provision to co-careering: Integrated guidance as a new approach to e-guidance in Norway. Journal of the National Institute for Career Education and Counselling, 41(1), 48-55. Barnes, A., La Gro, N., & Watts, A. G. (2010). Developing e-guidance competences: the outcomes of a two-year European project to transform the professional development of career guidance practitioners. Journal of the National Institute for Career Education and Counselling, 25(1), 26-32. Federal Reserve Bank of St. Louis (2023). FRED Economic Data (accessed May 5, 2023). https://fred.stlouisfed.org. Groysberg, B., Lee, J., Price, J., & Cheng, J. Y.-J. (2018). The leader’s guide to corporate culture: how to manage the eight critical elements of organizational life. Harvard Business Review, 96(1), 44-52. GEGS Peer Expert Group (2023). Different regions, common (good-e-guidance) stories. MetropolisNet – European network of local development partnerships in metropolitan areas. Haenlein, M., & Kaplan, A. (2019). A brief history of Artificial Intelligence: on the past, present, and future of Artificial Intelligence. California Management Review, 61(4), 5-14. Huang, M.-H, Rust, R., & Maksimovic, V. (2019). The feeling economy: managing in the next generation of Artificial Intelligence (AI). California Management Review, 61(4), 43-65. ILO (2021). Investing in career guidance (accessed 8 August, 2023). ILO.org. Koch, J. L. (1974). Employing the disadvantaged: lessons from the past decade. California Management Review, 17(1), 68-77. Krishna, R., Lee, D., Fei-Fei, L., & Bernstein, M. S. (2022). Socially situated artificial intelligence enables learning from human interaction. PNAS 119(39), e2115730119. Moore, D. A. (2017). How to improve the accuracy and reduce the cost of personnel selection. California Management Review, 60(1), 8-17. Paul, K. I., & Moser, K. (2009). Unemployment impairs mental health: meta-analyses. Journal of Vocational Behavior, 74(3), 264-282. Pentina, I., & Tarafdar, M. (2014). From “information” to “knowing”: exploring the role of social media in contemporary news consumption. Computers in Human Behavior, 35, 211-223. Schroth, H. Are you ready for Gen Z in the workplace? California Management Review, 61(3), 5-18. Tambe, P., Cappelli, P., & Yakubovich, V. (2019). Artificial Intelligence in human resources management: challenges and a path forward. California Management Review, 61(4), 15-42. Twenge, J. M. (2017). iGen: Why today’s super-connected kids are growing up less rebellious, more tolerant, less happy – and completely unprepared for adulthood – and What That Means for the Rest of Us. Atria Books. U.S. Bureau of Labor Statistics (2023). Labor force statistics from the current population survey (accessed May 5, 2023). bls.gov. Windley, D. (2022). Is AI the solution to hiring bias (or the cause of it)? (accessed August 3, 2023). Forbes.

Corporate interest in preserving nature and biodiversity is at an inflection point. Our analysis of earnings calls of 2,000 companies found that the percentage of companies mentioning “biodiversity” rose from 1.5% in 2017 to 10.5% in 2022. These numbers are only set to increase. In our recent survey of 1,500 business executives, 71% stated that their business success depends on the preservation of natural ecosystems. 84% believe that biodiversity loss poses a significant threat to their business, and more than half said that their understanding of how biodiversity loss impacts their organization has increased over the last three years. Apart from other factors, governments and regulations also have a lot to do with the increasing focus on biodiversity. Two recent developments have ensured that businesses now put nature and biodiversity on par with climate or net-zero goals: The acceptance of the 30x30 goal at COP15 (committing to the protection of 30% of land and 30% of coastal and marine areas by 2030) and the UN High Seas Treaty (offering protection to the vast tracts of ocean that are beyond national control). Detailed interviews with more than 24 senior executives revealed that a transition is under way, with biodiversity moving from a “for good” initiative to being a core part of the business. However, the executives also stated that the transition is far from smooth. This view is supported by our survey results, which show that only 19% of executives strongly agree that the private sector is doing enough to protect biodiversity. There are numerous challenges companies need to overcome to identify their nature-related interdependencies and act on them, one of them being scarcity of information. Technology can bridge this knowledge gap. AI-enabled systems can apprehend and make sense of underlying patterns too complex for the unaugmented human brain. Armed with this new knowledge, these systems can potentially predict where biodiversity loss might cause supply chain interruptions or cost shocks for maintaining plant and property. But companies appear to be confident about deploying technology for biodiversity—only one in four consider it a barrier. More than 91% of companies think digital technologies will become important in the next three years for managing nature- and biodiversity-related risks and opportunities. The catch is that these technologies are data-dependent, while the extent, quality, aggregation, use and measurement of nature and biodiversity data vary enormously between and within organizations. Executives are increasingly facing practical questions such as: How can we get better quality data for sensing our environment? How can we aggregate data from multiple sources—sensors, satellite, drones—for quick analysis? How can we better package insights for key decision-makers? How can we measure the impact of conservation decisions? Generating value with nature (and technology) More than ever, companies are using AI to advance their biodiversity initiatives. AI is useful not only for monitoring and conserving ecosystems but also for making more informed and specific business decisions. For example, AI-based tools are currently used to map ecosystem changes or to integrate data across different business units (see Figure 1). Figure 1. AI for biodiversity Source: Accenture survey, “Power business growth by protecting biodiversity”, April-May 2023.   Potential use cases of advances in AI, such as generative AI, are still waiting to be explored. For example, generative AI can produce synthetic data based on small biodiversity datasets. This synthetic data can then potentially be used to create traditional AI platforms or sophisticated digital twins for biodiversity preservation and restoration. The BioDT project, funded by the European Union, is currently developing a digital twin prototype for advanced modelling, simulations and predictions for evidence-based biodiversity restoration. And this is only the beginning. Companies using analytic tools and technologies to fulfill bold commitments for preserving nature and biodiversity are seeing clear benefits. For instance, Dow, in pursuit of its Valuing Nature Goal, collaborated with The Nature Conservancy (TNC) and the EcoMetrix Solutions Group to develop and deploy financial modeling tools to accurately put a value on nature. The Valuing Nature Goal is based on the premise that real business benefits come from fully informed decisions about how its operations rely on and impact nature. The company aims to achieve US$1 billion in net present value through projects that benefit both ecosystems and business as a part of its 2025 sustainability goals. The first big breakthrough of the collaboration was the Ecosystem Services Identification and Inventory (ESII) tool, a free app that allows users to quickly generate information on the ecosystem services performed by a specified landscape. TNC also helped Dow integrate consistent data on the environmental conditions on all of their global sites and then condense the data into a “Nature Scorecard.” The scorecard, combined with the ESII tool, allows Dow to compare the financial and environmental performance of potential nature-based interventions. For example, it helped Dow understand that if it restored forest, prairie and wetlands, it could save $2 million over 10 years by reducing mowing, fencing and stormwater management. Dow will use its analytic tools to identify cost-effective, nature-positive solutions to address coastal climate risk and water reliability issues around its most water-stressed sites. The ESII Tool and Nature Scorecard stand alongside traditional financial metrics in evaluating project performance. Now that Dow can value nature, it can generate value with nature. Towards intelligent biodiversity preservation While the value of biodiversity preservation and the return on investment can be high, companies still need to overcome challenges in their journey towards being nature-positive and biodiversity-positive. We suggest three AI-infused actions to strengthen their biodiversity initiatives.   1. Generate dependable data Companies can use a variety of data sources to monitor and gain insights into ecosystems and biodiversity, such as IoT, drones, satellite imagery and computer vision or acoustics. However, the validity and certifiability of this data is often a concern. For example, data from IoT sensors—such as temperature, humidity, soil quality, animal movements and plant growth—can be useful for biodiversity conservation, but is not always reliable. Sensors deployed in wild or harsh environments are prone to component failure and interference. Verified Telemetry is overcoming these challenges through automated monitoring of electrical properties of sensors and providing alerts for action. Jaljeevika, a non-profit that aims to improve the livelihood of small-scale fish farmers, is using dependable IoT technology in partnership with Microsoft and Accenture to ensure decisions are based on reliable data. It uses a range of IoT devices that monitor temperature, total dissolved solids and pH of water bodies. Farmers rely on these sensors deployed in remote locations for reliable data and advice. Microsoft Research’s dependable IoT technology was employed to rigorously test the advisory framework’s reliability, as well as data integrity and accuracy. Using Microsoft Azure IoT Hub and PowerApps, the solution incorporates heuristic models to accurately calculate ammonia levels in ponds, eliminating the need for ammonia sensors. This enables aqua farmers to optimize feed and enhance fish yield, while predictive insights and expert advisory bulletins simultaneously reduce operational costs. 2. Create analysis-ready datasets Effective use of biodiversity data requires integration of diverse data sources, including asset, supply chain, observation and financial data, which often comes in different formats and lacks standardization. One in three business leaders says that aggregating data, metrics and reporting at the organization level is one of the biggest challenges in biodiversity technology investments.  The lack of data infrastructure, ownership, interoperability and data security further complicates the process of deriving actionable insights. For instance, despite recent advances, the full potential of Earth observation data for biodiversity assessment, monitoring and conservation remains largely untapped. To overcome these challenges, companies need to invest in robust data infrastructure, establish data governance policies and leverage advanced technologies to automate data processing and create analysis- ready data sets. But while companies are building analysis-ready datasets, the data-centric AI movement allows companies to use limited datasets effectively and achieve AI business value and full-scale production of AI projects. Landing AI’s LandingLens visual inspection platform, which can potentially be used in various nature and biodiversity projects, applies computer vision to create models for image classification, object detection and image segmentation. Image segmentation can distinguish between different objects or classes based on pixel similarities. For example, it can distinguish between soil and mulch when companies need to remotely monitor and verify if suppliers are complying with regenerative agriculture practices. The neural network model is trained with satellite and unmanned aerial vehicle images that have labels for the objects of interest. It can be trained with visual prompts from end-users—as few as dozens of images —to achieve reasonable precision and recall for enterprise applications. 3. Put data into action AI, analytics and visualization tools can bring to life data aggregated from multiple sources, helping companies capture the nuances of species interactions, ecosystem dynamics and conservation challenges for informed decisions and targeted strategies. Intelligent platforms such as Microsoft Premonition can identify and monitor both invertebrate and vertebrate species using genomic signatures. Premonition metagenomics AI helps researchers analyze genomic data and find potential new pathogens in the environment, animals and in clinical settings. It can also help take targeted action to protect and restore local ecosystems. Its design enables it to combine different types of data (such as real-time sensor data and genomic data) into a knowledge graph that can be tracked and updated for large-scale analysis in the cloud. Powerful innovation is often combinatorial. When paired with advances in synthetic biology, AI has the power to open entirely new revenue streams for forward-thinking businesses. Take the case of Basecamp Research, a biodiversity geospatial intelligence company that has built a vast protein code database by collecting samples from over 40 global expeditions. Using this data, it is applying AI to discover novel proteins that provide R&D teams with more starting points and test the candidates most likely to succeed. Basecamp funnels part of its profits into preserving biodiversity. Chemix, a California based start-up, has fine-tuned a deep learning model to discover and develop materials that can replace nickel, cobalt or lithium, metals that are mined, for manufacturing sustainable and efficient electric-vehicle batteries. Meta AI has created a generative AI model that predicts protein structures that may help restore ecosystems, produce cleaner energy and cure diseases. It trained a large language model to learn evolutionary patterns and generate accurate protein structure predictions that companies can incorporate into entirely new raw materials, products and services. Conclusion The exponential improvement in intelligent technologies has the potential to significantly bridge the gap between awareness and action to prevent biodiversity loss. However, it is crucial that companies prepare themselves to embrace these technologies, even as they continue to evolve. The very first step in this journey is to automate data collection and aggregation, followed by putting in place intelligent systems that can help plan and execute actions. The preservation of our precious ecosystems and life itself warrants the use of technology. The rewards far outweigh the challenges and risks.

If you’ve been following the news about generative AI, you won’t be surprised that people are starting to work with large language models (LLMs) in ways that seem like human-AI partnerships. It’s become common, for example, for people to start writing by prompting ChatGPT for a rough first draft. But this is just one of a range of ways a person might collaborate with an AI partner, as shown in Figure 1. Figure 1: Continuum of Human AI Collaboration In late 2022, CNET started using generative AI to produce news articles that humans edited into final copy.2 We were curious about how articles produced this way differed from human-written CNET articles3 in how they might appeal to human readers. It wasn’t so much content that we were curious about, however. Many have noted that ChatGPT and other similar models can tend toward making up facts, giving weird responses, even “hallucinating,” which at least partially explains the need for human editors. What we wanted to explore is whether a piece primarily written by an AI would resemble human-written articles in the more stylistic ways that generate subjective impressions in human readers. This, we reasoned, has also to do with the shape of a piece of writing, not just its content. The idea that writing has a shape goes back to Aristotle. In Poetics, he argued that a good story has a beginning, a middle, and an end. By “beginning” he meant a written expression by the author that the reader would come to prepared already to recognize and understand, without any preparation by the author, and that would move the story toward subsequent issues. A beginning, then, has no antecedents, but does generate implications: new issues that need to be dealt with. Those next new issues to be dealt with are middles. A middle takes up issues that flow from beginnings and progresses them further. Middles don’t resolve the issues; they might even raise more of them. A middle, then, is both implied and has implications. Middles flow toward ends. And ends have the characteristic that they follow naturally from middles, but they do resolve issues, achieving some kind of closure. An end has antecedents (middles), but no succedents. Aristotle was pointing out something about how a good piece of writing hangs together, how its pieces interrelate to create a sense of unity. Or, as he put it: “the structural union of the parts being such that, if any one of them is displaced or removed, the whole will be disjointed and disturbed.”4 Plot, he said, is the soul of a good story. Many writers have described “shapes” in stories, such as Joseph Campbell’s “hero’s journey”5 or Kurt Vonnegut’s “8 shapes” of narratives (e.g., “Man in a Hole” and “Girl Meets Boy”).6 But news articles aren’t exactly stories and don’t, usually, have shapes like them. There is a journalistic plot that starts with “The Lead” (spelled “lede” by many journalists—a provocative hook with essential facts), moves to “The Body” (background details), and then to “The Tail” (extra info that adds richness). There might be similarities between the shapes of stories and news articles written by humans and AI, but there might also be differences. We examined this question empirically. Psychologists Ryan Boyd, Kate Blackburn, and James Pennebaker have developed a way of accessing the shape of a piece of writing based on the frequency of certain kinds of words at different points throughout the piece. They use “narrative arc analysis” accomplished by software called “Linguistic Inquiry and Word Count” (LIWC)7 to show that the early parts of most stories contain a lot of “words that pertain to nouns and how they [relate] to one another,” which Boyd et al. label “Staging language.” As they move to the middle of a story, authors “use more words that signal action,” which Boyd et al. call “Plot Progression language.” Moving further through a story, an author employs words that signal tension and conflict, and uncertainty about whether a character’s goals will be reached, from which Boyd et al. discern the level of “Cognitive Tension.” These measures provide a quantitative, graphical description of what the authors call a “narrative arc,” a reasonable approximation of what we have called plot. They find consistency in narrative arc when they analyze many stories from films, novels, short stories, etc. (see Figure 2). Figure 2: Narrative Arc Analysis of stories from Boyd et al. 2020 Narrative arc analysis does not “read” a story in the way that humans do. The LIWC software measures the frequency of words of different types. Staging language is determined by the frequency of prepositions and articles. Plot progression is measured by counting pronouns, auxiliary verbs, and other function words, and so on. Also, ChatGPT and its relatives are not “writing” the way humans do. LLMs generate structures in natural language that are like, but not a replica of, training data. The creators of ChatGPT (OpenAI) tell us that its LLM has trained primarily on a large corpus of available data through 2021 and extracts from that a sense of what words more appropriately follow the words it has already committed to. Figure 3 shows the narrative arc analysis for 77 AI-written articles from the CNET platform that were edited and fact-checked by two human editors.8 We also analyzed the narrative arc of AI-written articles reviewed by each of the human editors separately (see Figures 4a and 4b), and of a representative sample of CNET posts written by additional human authors (not just the two who edited the AI posts) during a period parallel to the composition of the AI-written posts (Figure 5). Figure 3 Narrative Arc Analysis for CNET posts written by AI, edited by humans Figure 4a Narrative Arc Analysis for CNET posts written by AI, edited by human editor #1 (n=27) Figure 4b Narrative Arc Analysis for CNET posts written by AI, edited by human editor #2 (n=50) Figure 5 Narrative Arc Analysis for all CNET posts written by all human authors during an interval parallel to the composition of AI-written articles There are not huge differences across these plots. Staging roughly decreases as the article progresses, as we might expect, as The Lead (the provocative hook) progresses to The Body (background details) and The Tail (additional info). Plot Progression shows some variation but generally increases through the article. Cognitive Tension increases too, though there is an interesting dip at the end in the graph for all human writers, perhaps suggesting that human writers are more inclined toward achieving closure at the end, as in a story narrative. CNET articles need not have a similar structure to that of a film, novel, or short story. They are different genres of writing. But it is interesting that there seems to be in news articles an echo of the shape of stories, perhaps indicating a shape within writing that transcends genre. It makes sense that you’d need to do more Staging early in any kind of written work, and that writers might use Plot Progression as an attention-keeping mechanism, regardless of genre. It makes sense, too, that Cognitive Tension is not the same for a CNET’s article as in a classic narrative story. Maybe a news article doesn’t achieve closure at the end because there’s an aim to make you want to come back or read the next story. Whereas, in a classic film or play, the audience will be unhappy if the story doesn’t reach closure. From the similarities in these graphs, we might conclude that AI is up to the task of generating acceptable plots for news articles, with a human editor at least. In writing articles that are then edited by humans, then, an AI-partner seems to perform competently. But can it be an equal partner in a creative collaboration? LLMs as Equal Partners in Symbiotic, Creative Collaboration Researchers agree that to be considered creative an outcome must be original (unlike outcomes that have preceded it) and useful (or, more generally, valuable – usefulness is not the only kind of value).9 People and organizations, however, struggle to generate original ideas and recognize value in unfamiliar outcomes.10 We tend to cling to past ideas and old ways of assessing what is valuable. How might an LLM contribute to breaking these patterns as part of an innovative partnership? One might presume that since an LLM derives its responses entirely from data about the past, its outputs would be derivative rather than original. However, Professor Ethan Mollick of Wharton Business School has found that LLMs can make surprising connections that yield unusual, seemingly non-intuitive, juxtapositions.11 A glance at the images DALL-E can generate will disabuse you of the idea that nothing original can come of inferences from training data. Mollick demonstrates how we can use LLMs to get unstuck from a too-narrow set of ideas. Continuous experimentation with such tools, in which you start with a prompt, receive a result, and then build on that by asking subsequent prompts, exemplifies Symbiotic AI to at least some degree. In great creative collaboration, however, you might well want more than this from a partner. We draw on here the idea of “ensemble,” a notion from the world of the arts. In an ensemble, each collaborator makes fresh choices that, when shared, provoke additional fresh choices, which, in turn, provoke additional fresh choices. This chain reaction continues among collaborators until they are “successful in combining their voices into a coherent whole…[and] individual members relinquish sovereignty over their work and thus create something none could have made alone.”12 In a string quartet, for example: “[T]he relentless sequence of new actions and reactions…[creates] an intense and joyful  experience for [collaborators who are] confident in their abilities to solve the problems that [come] at them… Leadership passes from member to member, from moment to moment… All of this depends on a high standard of listening to each other… The quality of communication in the ensemble [grows] so high that members [can] sense when another member [is] about to make a mistake, and even when he recover[s] and avoid[s] it.” 13 We can see elements of this kind of symbiosis in the iterative interactions that Mollick describes, but not at the level apparent in this description. While LLMs can be helpful collaborators, they currently fall short of achieving the highest levels of symbiosis. Implications for Business Our analysis suggests several actions that managers should take to prepare their organizations to maximize the potential of creative collaboration with generative AI. Pay attention to the “shape” of work generated by AI, not just the content – How receptive human audiences are to AI-generated content is not just a matter of content, but also factors like narrative arc or plot. The mainstream conversation about generative AI has been very content focused. Get ahead of this by promoting an awareness within your organization that it matters how messages are conveyed, not just what they say. Use automated tools to help monitor the shape of AI-generated content – Evaluating creative works, such as writing, has often largely relied on human judgment. But there are now analytics-based techniques, such as narrative arc analysis, that can help check if AI-generated content conforms to human writing conventions for specific genres. But maintain human editorial oversight – Just because AI appears to be pretty good at generating the narrative structures of human writing genres does not mean that you should let AI generate content unsupervised. It is well documented, of course, that AI-generated content can contain factual errors14 or exhibit bias.15 But you might also want to monitor for differences in shape, in the inclination to achieve some closure at the end of an article, for example, to make sure your AI-generated content is palatable for human audiences. Develop methods for more symbiotic AI collaboration, that move toward ensemble – Generative AI tools are useful in creative collaboration, but they do not yet achieve the highest levels of symbiosis, such as ensemble. Ensemble occurs only when individual collaborators make fresh choices that provoke further fresh choices from others in a highly dynamic chain reaction. Creative collaboration with AI partners will require, then, development of increasingly dynamic processes of interaction between human and AI models. The iterative, prompt-based nature of LLMs is already conducive to this, but there will likely be a learning curve involved in moving toward true ensemble. Make moving down this learning curve a formal project within your organization. Monitor stylistic trends in your AI-generated content – Today, LLMs train predominantly on human-generated content. But as more LLMs are released “into the wild,” they will increasingly train also on AI-generated content. Research suggests that this might affect the quality and shape of the content.16 Differences between AI and human-generated content may become amplified, resulting in narrative shapes that start to diverge from what is natural for humans. This could generate new levels of creativity. Or it could begin to generate content that humans experience as strange or unappealing. Organizations engaged in creative collaboration with AI will need to keep an eye on this. AI is becoming capable of performing tasks that were once exclusive to humans. This does not diminish the importance of human activity but rather brings new meaning to it. It is important for today’s managers to be thinking about this now, to prepare their people for new roles and their organizations for an increasingly exciting future with new kinds of creative partnerships. References The idea of augmenting human intellectual capabilities by using computers has a long and distinguished history. See Douglas Engelbart’s seminal 1962 paper, “Augmenting Human Intellect: A Conceptual Framework,” Englebart, accessed July 4, 2023. Harrington, C. (2023). CNET Published AI-Generated Stories. Then Its Staff Pushed Back, May 16, CNET, accessed July 11, 2023. When they were first written, CNET’s AI-written articles were clearly labeled on their website as AI-generated (see, for example, CNET), which allowed us to easily distinguish between human- and AI-written pieces. We collected data in February 2023, when articles were labelled as written by CNET Money and edited by human editor. However, in June, CNET overhauled its AI policy and updated past stories (The Verge). These changes make it harder, now, to distinguish between articles that are AI- and human-generated. The same URL that once took you to an article clearly identified as AI-generated might today take you to a modified version of that same article that shows a human byline. For those interested, early versions labeled “AI-generated” can be independently recovered and verified from the Wayback Machine web archive. Aristotle. (1895). Poetics. trans. S. H. Butcher, no longer under copyright, p. 16 in original edition. Campbell, J. (1990). The Hero’s Journey. New World Library. Johnson, S. (2022). Kurt Vonnegut on the 8 “shapes” of stories. [Big Think] (https://bigthink.com/high-culture/vonnegut-shapes/), accessed July 8, 2023; see also Frye, N. (1957). Anatomy of Criticism. Princeton University Press. Boyd, R.L., Blackburn, K.G. & Pennebaker, J.W. (2020). The narrative arc: Revealing core narrative structures through text analysis. Science Advances, 6(32), p.eaba2196. The narrative arc analysis requires that we specify the number of “segments” (or parts) we want to divide an article into for the purposes of the analysis. The prior literature recommends five segments for this analysis. Although we present our analysis for five segments (represented on x-axes of Figures 3, 4, and 5), we have also performed this analysis for different number of segments and found similar results. Amabile, T.M. (2018). Creativity in Context: Update to the social psychology of creativity. Routledge. Simonton, D.K. (1999). Creativity as blind variation and selective retention: Is the creative process Darwinian? Psychological Inquiry, pp.309-328. Mollick, E. (2023). One Useful Thing, oneusefulthing, accessed June 23, 2023. Austin, R. D. & Devin, L. (2003). Artful Making. Financial Times Prentice Hall, p. 16. Austin, R. D. & O’Donnell, S. (2007). Paul Robertson and the Medici String Quartet. Harvard Business School case 607-083, p. 8; see also, Peat, F. D. (2000) The Blackwinged Night: Creativity in Nature and the Mind. Basic Books, p. 191. Sato, M. & Roth, E. (2023). CNET found errors in more than half of its AI-written stories. The Verge, January 25, The Verge, accessed May 28, 2023. Shanklin, W. (2023) AI Seinfeld was surreal fun until it called being trans an illness. engadget, February 6, [Engadget] (https://www.engadget.com/ai-seinfeld-bigoted-transphobic-hateful-moderation-193449772.html), accessed May 28, 2023. See Shumailov, I., Shumaylov, Z., Zhao, Y., Gal, Y., Papernot, N., & Anderson, R. (2023). The Curse of Recursion: Training on Generated Data Makes Models Forget. ArXiv, abs/2305.17493.

Generational labels like ‘Baby Boomer’, ‘Gen Z’ and ‘Millennial’ make for seductive clickbait. News feeds are flooded with research articles, editorials and whitepapers that claim to unlock the mysteries of what each generation knows and wants. Meanwhile, social media dishes up satirical videos that stereotype how each generation ‘shows up’ (or doesn’t) at work. But, there is growing concern about the negative influence that these cohort labels can have on workplace attitudes, in addition to the lack of scientific rigour that underpins them.1 A report from the US National Academies of Sciences, Engineering, and Medicine questioned the use of generational labels in the workplace, fearing that many generational studies lack rigour and give oxygen to age-related stereotypes that reinforce discriminative practices. Our view, as behavioural scientists, is that we should use generational labels responsibly and do more to understand when and how generations matter. As it stands, there is no official taxonomy or oversight committee who decide when a new generation starts or ends, or what to name it. Although these labels often evolve organically through popular media and public discourse, many originate from the Pew Research Center, a US Think Tank. Pew defines Generation Z as those born between 1997 and 2004, Generation Y or Millennials as those born 1981-1996, Generation X as those born 1965-1980, and Baby Boomers as those born 1952-1964. Pew has been facing mounting pressure to end their use of generational labels. Last month, they clarified how they will (and will not) use these labels in future research, while encouraging their readers to bring “a healthy dose of skepticism to the generational discussions”. Do Generational Labels Matter? Given the widespread use of generational labels, we set out to understand what they meant to workers. Surveying 1,500 workers across the UK and US using Pew definitions, nearly 60% identified with the generation that they belong to. Within the sample, ‘Baby Boomers’ were most accurate, with 81% identifying their generation, as compared to 51% of the ‘Gen Z’ sample (Figure 1). Many workers thought of themselves as belonging to one of the generations on either side of their own. For example, 9% of Millennials self-identified as Gen Z, while 7% identified as Gen X. This raises the inherent challenges of arbitrary generational cut-off points. Generational labels seem to have much stronger salience in the US than the UK, with only 15% of US workers not knowing which generation they belonged to, compared to 40% in the UK. Figure 1: Percentage of workers who correctly identify their Generation. One-third of the sample also felt that the generation that they belonged to was important to their personal identity. This importance was greater for older workers than for younger workers (Figure 2). Thus, not only are generational labels widespread, for many workers they are meaningful to how they see themselves and their place in the world. Figure 2: Importance of Generation to worker’s identity (by Generation). Moving Towards ‘Generational Thinking’  Labels aside, ‘generations’ and ‘generational thinking’ are important concepts that can be beneficial when used appropriately. Given that we are living and ,therefore, working longer, most major firms will have five generations working side by side. But regrettably, the productivity benefit of generational diversity often goes unrealised, with generations failing to relate to each other. For example, those from younger generations are sometimes unfairly labelled as ‘snowflakes’ for placing greater importance on diversity, equity, and inclusion (DEI),  corporate social responsibility (CSR), and work-life balance than those who came before them.2,3,4 Meanwhile older generations can be unfairly stereotyped as resistant to change or lower performing, leaving older talent undervalued or cast aside too early.5,6 These misunderstandings can create tensions that prevent successful intergenerational inclusion and collaboration.7 It is hard to deny that differences between age cohorts exist when the era that people were born into will have invariably shaped their shared experiences of history. For example, the COVID-19 pandemic will have affected people differently based on multiple factors, including their country of residence, their financial status and, of course, their age.8 Similarly, workplace learning and communication preferences can be influenced by the technologies available to each generation during formative years.9 Navigating Generations at Work Generational labels are widespread and widely understood. But rather than rely too heavily on them as a ‘one size fits all’ notion, management decisions about attracting, developing, and retaining talent are better guided by an understanding of ‘generational thinking that recognises and responds to the challenges of an increasingly intergenerational workforce. The challenge ahead then becomes how to acknowledge the importance of generational experience and diversity, while avoiding unhelpful stereotypes and subsequent discrimination. Here, we offer 4 ‘do’s and don’ts’ that will help to this end: Do recognise the value of intergenerational diversity. Complex problems demand a wide variety of perspectives. Generational diversity is particularly important to tackling complex business problems as it delivers the knowledge, perspectives and connections to increase creativity, innovation, and performance.10,11,12 Sales and service benefit from teams that reflect the generational diversity of their customers, and working in an organisation where people of different ages enjoy positive relationships, free from discrimination, is good for the job satisfaction of every generation.13,14 Don’t use generational labels to reinforce negative stereotypes. Generational labels are socially widespread. There are situations where cohorts share similar experiences or attitudes that are influenced by the era that they were born into. But, as we have shown, there are no absolutes and individuals are not definitively defined by their generation. Labels can be inexact and can veer into unhelpful stereotypes and pseudoscience. They can also be hurtful when they become discriminatory. Creating a team culture that discourages generational divisions and promotes generational diversity drives increased performance.15 Do acknowledge that generational differences exist… Social and economic influences can affect different generational groups differently. For example, degree-educated Gen X’s who graduated university and entered the job market in the 1990s, will have likely had qualitatively different experiences to degree-educated Millennials entering the market post the Global Financial Crisis. For some, these experiences and the labels that accompany them might speak to directly to their experience and how they want to be understood. …But don’t overstate these differences. Our findings revealed that most workers do not consider their generation to be overly important to who they are as a person, suggesting that managers must not ignore individual experiences by focusing too much on someone’s generational category. And rather than attempt to coin provocative new categories, like ‘Geriatric Millennials’, we must distinguish generational experiences from societal shifts that affect everyone (e.g., increased mobile phone use), or age-related trends that occur at different stages of the life course (e.g., ageing; developing careers; relational changes).  The popularity of generational labels cannot be expected to fade anytime soon. The categorizations provide greater nuance than simply referring to ‘older and younger’ workers and are less arbitrary than referring to groups based on their age bracket. But rather than focusing too heavily on labels, managers can move towards ‘generational thinking’ by recognising the value of intergenerational teams and avoiding stereotypes. This means questioning if the characteristics we observe might persist or change in groups as they age, or if these characteristics reflect broader changes that affect workers of all generations. References Rudolph, C. W., Rauvola, R. S., & Zacher, H. (2018). Leadership and generations at work: A critical review. The Leadership Quarterly, 29(1), 44-57. https://doi.org/10.1016/j.leaqua.2017.09.004 Schroth, H. (2019). Are you ready for Gen Z in the workplace? California Management Review, 61(3), 5-18. cmr.sagepub.com. Singh, R., Chaudhuri, S., Sihag, P., & Shuck, B. (2023). Unpacking generation Y’s engagement using employee experience as the lens: an integrative literature review. Human Resource Development International, 1-29. Human Resource Development International. Sánchez-Hernández, M. I., González-López, Ó. R., Buenadicha-Mateos, M., & Tato-Jiménez, J. L. (2019). Work-life balance in great companies and pending issues for engaging new generations at work. International journal of environmental research and public health, 16(24), 5122. International Journal of Environmental Research and Public Health. Ng, T. W., & Feldman, D. C. (2012). Evaluating six common stereotypes about older workers with meta‐analytical data. Personnel psychology, 65(4), 821-858. Personnel Psychology. Ng, T. W. H., & Feldman, D. C. (2008). The relationship of age to ten dimensions of job performance. Journal of Applied Psychology, 93(2), 392–423. Journal of Applied Psychology. North, M. S., & Fiske, S. T. (2015). Intergenerational resource tensions in the workplace and beyond: Individual, interpersonal, institutional, international. Research in Organizational Behavior, 35, 159-179. Research in Organizational Behavior. Hamilton, O. S., Cadar, D., & Steptoe, A. (2021). Systemic inflammation and emotional responses during the COVID-19 pandemic. Translational Psychiatry, 11(1), 626. Translational Psychiatry. Lowell, V. L., & Morris Jr, J. (2019). Leading changes to professional training in the multigenerational office: Generational attitudes and preferences toward learning and technology. Performance Improvement Quarterly, 32(2), 111-135. Performance Improvement Quarterly. Salas, E., Reyes, D. L., & McDaniel, S. H. (2018). The science of teamwork: Progress, reflections, and the road ahead. American Psychologist, 73(4), 593–600. American Psychologist. Wegge, J., Jungmann, F., Liebermann, S., Shemla, M., Ries, B. C., Diestel, S., & Schmidt, K. H. (2012). What makes age diverse teams effective? Results from a six-year research program. Work, 41(Supplement 1), 5145-5151. IOS Press Li, Y., Gong, Y., Burmeister, A., Wang, M., Alterman, V., Alonso, A., & Robinson, S. (2021). Leveraging age diversity for organizational performance: An intellectual capital perspective. Journal of Applied Psychology, 106(1), 71–91. Journal of Applied Psychology Cavusgil, E., Yayla, S., Kutlubay, O. C., & Yeniyurt, S. (2022). The impact of demographic similarity on customers in a service setting. Journal of Business Research, 139, 145-160. Journal of Business Research. King, S. P., & Bryant, F. B. (2017). The Workplace Intergenerational Climate Scale (WICS): A self‐report instrument measuring ageism in the workplace. Journal of Organizational behavior, 38(1), 124-151. Journal of Organizational Behavior. Homan, A. C. (2019). Dealing with diversity in workgroups: Preventing problems and promoting potential. Social and Personality Psychology Compass, 13(5), e12465. Social and Personality Psychology Compass.