AI in Business Contexts - Human-AI Teams PDF

AI in Business contexts Human - AI teams PGE M1 Nabila BOUKEF Ph.D. Associate Professor in information Systems and digital transformation SKEMA Business School ©Boukef, 2024-2025  Questions addressed...

AI in Business contexts Human - AI teams PGE M1 Nabila BOUKEF Ph.D. Associate Professor in information Systems and digital transformation SKEMA Business School ©Boukef, 2024-2025  Questions addressed in this lecture 1. What are the characteristics of Human-AI teams? 2. Is human creativity still important in a world shaped by generative AI? 3. How can managers best harness human-AI collaboration to produce novel and valuable outcomes, while preserving the human skills crucial for idea generation? Course content Course structure Introduction 1. Human - AI teams: Opportunities and challenges 2. Human - AI team creativity 3. Human - AI team creativity and the future of work Conclusion Introduction The need to understand human-AI teams “The greater the degree of organizational focus on people helping AI, and AI helping people, the greater the value achieved.” (Accenture’s CEO, 2020, HBR,2020) “The greatest impact of intelligent technologies won’t be from eliminating jobs but from changing what people do and driving innovation deeper into the business.” (Barro and Davenport, 2019, MIT SMR) “We are now solving problems with machine learning and artificial intelligence that were in the realm of science fiction for the last several decades”. Jeff Bezos, former Chairman of Amazon (Raisch, 2022) To survive in the digital era, companies must innovate rapidly Creativity is essential to innovate. While creativity is about coming up with ideas, innovation is about making those creative ideas happen. Creativity involves generating ideas that are both novel and useful, such as proposing ways to enhance departmental performance, optimizing processes and routines, improving quality, and developing innovative solutions to emerging challenges. To remain competitive, companies must foster an environment that encourages creativity and enhances employee innovation. Creativity: Recombining Diverse Knowledge and Perspectives in Teams Team creativity can be defined as the outcome of the integration and combination of the creativity of team members (Pirola‐Merlo & Mann, 2004). Team creativity is a social process in which relationships among team members play a critical role (Taggar, 2002). Like human-human teams, team creativity in human-AI teams depends on the creativity of the different members and results from the combination of the different knowledge. 1. Human - AI teams: Opportunities and challenges Definition and characteristics of human-AI teams A human-AI team is defined as comprising at least one human and one artificial agent, where AI and humans together form a "coupled system" in which each contributes unique expertise, fostering a collaborative network for decision-making. Characteristics of a Human-AI team  members are interdependent, Image generated by Dall-E  share a common goal,  maintain a significant degree of independence. (O’Neill et al., 2020).  Conditions for Synergy in Human-AI Teams: The National Academies of Sciences, Engineering, and Medicine (2021) defines four conditions for a human-AI team to profit from these synergies: (1) The human part has to be able to understand and anticipate the behaviors of the deployed intelligent agents. (2) To ensure appropriate use of AI systems, the human should be able to establish an appropriate relationship of trust. (3) The human part can make accurate decisions when using the output information of the deployed systems and (4) has the ability to control and handle the systems appropriately. Berretta et al., 2023  Class activity: Reflect and discuss Work in pairs to identify specific examples of AI use within an organizational context. Discuss how these applications impact work, such as by improving efficiency, enabling faster decision-making, and other benefits  Human - AI team: Automation vs. Augmentation  Automation transfers tasks entirely to machines, minimizing or eliminating human input. It brings efficiency, speed, and consistency.  Up to 30 percent of hours worked could be automated by 2030, boosted by gen AI, leading to millions of required occupational transitions. https://www.mckinsey.com.  Augmentation allows employees to leverage AI to work smarter, combining human judgment with digital tools for better outcomes. Unlike automation, augmentation emphasizes ongoing human- machine interaction and encourages experimentation.  The decision between augmentation and automation hinges on the task's nature: routine, well- structured tasks are suitable for automation, whereas complex, ambiguous tasks are better suited for augmentation.  Examples of automation and augmentation Customer Service Chatbots: Healthcare Diagnosis: Augmentation Automation AI algorithms assist doctors by analyzing AI systems can automate responses to medical images, suggesting potential customer inquiries, handling routine diagnoses, or predicting treatment outcomes. questions and requests, while human While AI handles large-scale data analysis, agents step in for more complex issues. In this case, AI is primarily responsible humans provide critical judgment, contextual for automating the task of responding to understanding, and patient care that the AI customers. cannot replicate. Companies should consider AI’s potential to enhance human capabilities rather than replace them. Risks of Automation: While automation can free resources, it encourages short-term thinking, loss of expertise, and lock-in effects, making it harder to adopt new and innovative practices. Achieving synergy between automation and augmentation is essential for businesses to innovate, stay competitive, and ensure that technology empowers rather than replaces human roles. How do the team members work together? Research highlighted two options related to Human and AI collaboration: Tasks divisions (Tasks specialization): Humans and AI Perform separate tasks based on their relative advantages in performing these tasks. While specialization can enhance overall performance, when humans are responsible for only certain parts of a task rather than the entire process, they may lose the skills needed for the components they no longer engage with. Collaborative work (Ensembling): Humans and AI Work Together on the Same and then aggregate their outputs. Working together enables  Examples Hiring process: tasks Healthcare Diagnosis (Ensembling) specialization AI Analyzes medical images, detects AI can handle the initial sub-task of potential issues (e.g., tumors), and screening applications by reviewing suggests possible diagnoses based on resumes and shortlisting candidates. historical data. The human role then focuses on Human Reviews AI suggestions, conducts conducting in-depth interviews with the further tests, and makes the final shortlisted candidates. diagnosis, considering the patient's history and unique context. This division of tasks allows AI to streamline preliminary screening, while Collaboration: AI provides insights and humans perform the more nuanced recommendations, while humans use their evaluation expertise to interpret, validate, and make nuanced decisions.  Unlocking Human-AI Team Synergy: The Power of Delegation and the Metaknowledge Gap (Fugener When humans andet al., 2022) AI collaborate, they can achieve better results than AI working alone, even if the AI already outperforms humans individually. Collaboration between humans and AI leads to better results when AI delegates tasks to humans, but not when humans delegate to AI. Humans struggle with delegating tasks effectively to AI, even when they recognize the AI’s support, due to a lack of metaknowledge—incorrect self-assessment of their own capabilities. The challenge in human-AI collaboration stems from humans’ unconscious inability to accurately assess their abilities, which limits effective decision-making and collaboration with AI. Fügener, A., Grahl, J., Gupta, A., & Ketter, W. (2022). Cognitive challenges in human–artificial intelligence collaboration: Investigating the path toward productive delegation. Information Systems Research, 33(2), 678-696.  The role of human experience in Human-AI teaming (Wang, Gao and Agarwal, 2024)  The context of the research is the medical coding industry.  The study focuses on risk adjustment coding from medical charts, where workers review the complete chart, especially unstructured physician notes, to determine if the patient has certain medical conditions, such as diabetes and cardiac disease.  Analysis of data from 1,120,831 patients charts over one-year period.  To streamline the labor-intensive process of risk-adjusted medical chart coding a machine learning-based AI was developed. Results  The results show that AI benefits workers with greater task-based experience.  However, senior workers experience less benefit from AI compared to their junior counterparts.  The lower productivity increase among senior workers is not directly attributable to their seniority, but rather to a lower level of trust in AI, likely stemming from their broader job responsibilities. Wang, W., Gao, G., & Agarwal, R. (2024). Friend or foe? Teaming between artificial intelligence and workers with variation in experience. Management Science, 70(9),1 -23.  Skill-Biased Creativity (Jia et l.,2024)  A field experiment in a telemarketing company showed that AI assistance (Chat bot) in lead generation boosts employees' creativity in customer engagement, ultimately leading to higher sales.  Higher-skilled employees experience more creativity and positive emotions with AI support, developing innovative scripts, while lower- skilled employees show limited creativity gains and may feel negative emotions.  AI enhances creativity but predominantly benefits employees with greater expertise, highlighting a skill bias in achieving AI-augmented creative outcomes. Jia, N., Luo, X., Fang, Z., & Liao, C. (2024). When and how artificial intelligence augments employee creativity. Academy of Management Journal, 67(1), 5-32.  Takeaways 1: Automate/Augment Specialization/ Ensembling Gain in efficiency (increase in productivity, …) Expected outcomes (improved decision-making, improved creativity) Unexpected outcomes (loss of expertise,…) This relationship depends on many factors such as Human metaknowledge, experience with the task, trust in AI 2. Human - AI team creativity  Generative AI is gaining momentum Investment in generative AI is increasing (80% of companies have increased their investment in generative AI from last year). (Capgemini Research Institute survey, 2024) According to Gartner, over 80% of enterprises are expected to either utilize application programming interfaces (APIs) that grant access to generative AI capabilities or implement applications powered by generative AI by 2026. Around 25% of the total work time of knowledge workers is spent on the kinds of creative tasks that generative AI is beginning to get good at. (Gratton, 2023, MIT SMR).  What can Gen AI do? Improving Efficiency. Accelerating manual or repetitive tasks, such as writing emails, coding, or summarizing large documents. Generating Content and Ideas. Creating new, unique outputs across a range of modalities, such as a video advertisement or even a new protein with antimicrobial properties. It can also create art, music, and other forms of media, often mimicking human creativity. Personalizing Experiences. Creating content and information tailored to a specific audience, such as chatbots for a personalized customer experiences or targeted advertisements based on patterns in a specific customer's behavior. Generative features enable the model to engage in meaningful and dynamic conversations, responding to user inputs in a natural and human-like manner. https://www.bcg.com/capabilities/artificial-intelligence/generative-ai  What makes Gen AI so different from other AI tools? Generative and Conversational Features: The ability to generate and engage in conversations is revolutionary. Broader Access to AI Capabilities: These features make AI more accessible and easier to use than ever before. No specialized programming skills required. Higher expected Impact on Workforce: Studies estimate that GPT models will influence at least some tasks for about 80% of all workers. Gen AI can create new outputs from learned patterns as opposed to predictive AI, which anticipates future outcomes based on past learned patterns. Class activity: (20 min of creative work and then reflect on the process) Group work (Group of 4 students): Collaborative writing using Gen AI  You are part of a dynamic change management team at your company. Your task is to tell the story of your company’s digital transformation journey, bringing to life innovative changes and powerful successes along the way. 1- Start by writing a short paragraph or a few sentences that highlight compelling use cases. 2- invite the AI to help you shape the next part of the story. Along the way, you’re encouraged to include an image to visualize your story’s theme or main achievement—this could be anything from a concept illustration of your digital innovation to a picture representing the impact on your customers or teams.  Be creative and collaborative! Build on each other’s ideas, integrate the AI’s suggestions seamlessly, and create a cohesive, engaging narrative that captures the excitement and impact of your company’s transformation.  Class discussion What surprised you about using Gen AI in the creative process? Did the Gen AI help you think of ideas they might not have had otherwise? How did Gen AI contribute to or limit your creativity? Where do you see the balance between human and machine  The advancements with Gen AI pave the way for a new approach to creative problem-solving and ideation A new form of collaborative partnership between humans and AI How Gen AI boosts creativity? Gen AI can boost creativity by offering novel ideas and reducing the time spent on routine tasks. This leads to a rise in the quantity of creative contents. Empowering Creative Teams: Achieving More with Fewer Resources (using tools such as Udio and Suno AI enabled teams to create creative contents that used to require specialized teams and external agencies). Gen AI empowers creative content to adapt in real-time to user input and feedback, transforming static content into dynamic, personalized experiences that differ for each user (Example Cadbury).  GPT-4 Revolutionizes Idea Generation: Producing Hundreds of Ideas in Minutes with Superior Quantity and Quality (Girotra et al., 2023) The study is based on experiments involving students engaged in product design GPT-4 generates ideas that are perceived as more likely to be purchased than human- generated ideas, although these AI ideas are seen as high-quality, they are also viewed as less novel compared to human ideas. AI outshines human creativity in generating the highest-quality ideas, with AI-generated ideas being seven times more likely to rank in the top 10%. GPT-4 vastly outpaces human idea generation, producing hundreds of ideas in minutes, showing both quantity and quality benefits in ideation (Two hundred ideas can be generated by one human interacting with GPT-4 "in about 15 minutes. A human working alone can generate about five ideas in 15 minutes and humans working in groups do even worse). Girotra, K., Meincke, L., Terwiesch, C., & Ulrich, K. T. (2023). Ideas are dimes a dozen: Large language models for idea generation in innovation. Available at SSRN 4526071. Creativity Unleashed or Constrained There is a debate as to whether Gen AI will enhance creativity by providing new tools and inspiration, or whether it will impose a "creativity tax" by potentially stifling original thinking. The “creativity tax” refers to the idea that reliance on AI-generated content might reduce the originality and depth of human creativity: over-reliance on AI might lead to homogenized outputs and a decrease in unique, innovative thinking. Shifting control and responsibility for creativity to AI could potentially weaken individuals' confidence in their own creative abilities, reinforcing the (potentially self-fulfilling) belief that "machines are superior to humans in creativity. Castro et al. (2024) Takeaways 2: While generative AI offers opportunities to enhance creativity, Human AI-creativity task division can potentially hinder human creativity by limiting engagement with the full creative process. A collaborative approach is crucial, as over-reliance on AI may lead to homogenized results, stifling diversity of thought and innovation. Despite the capabilities of AI, human creativity remains essential, as it brings unique insights, emotional 3. Human - AI team creativity and the future of work Human - AI teams in an organizational context Organizational level Organizational context Human creativity AI features Use and interactio ns Team level Corporate Policy regarding the use of Gen AI Creativity at the team level (Human-AI team)  Generative AI features Users can have access to a set of Gen AI tools depends on their needs and the characteristics of the task they have to handle. Generative AI features : There is a growing number of Gen AI technologies with different features. Users are more likely to achieve higher benefits when the features of the Gen AI fit the needs of the task (Task-technology fit). Simultaneous use of generative AI tools with similar features may lead to overload and inefficiency, due to the risk of redundancy. Trust in the AI teammate is very important to get the most out of the collaboration between human and AI.  Communication with Gen AI: prompting  Like Human-human teams, communication is important to foster creativity. Iterative  Prompting is used to communication with Gen AI. process And continuous improvements Prompt  The importance of the iterative process. Chain of thoughts prompting: Answer Zero-shot : No structure given in the prompt. Gen AI resolves the issue on its own. One/few-shot: the prompt provides Step-by-step roadmap for Open AI to  The interaction with Gen A: iterative human prompting leads to more creative outcome (Boussieux et al., 2024) Human engagement with AI enhances output quality: The novelty and value of Human-AI outputs depend on how humans guide AI's search behavior, with iterative human prompts increasing both novelty and value. Human AI solutions were developed in significantly less time (5.5 hours) and at a lower cost ($27.01), offering a promising alternative to traditional creative problem-solving approaches. Boussioux, L., Lane, J. N., Zhang, M., Jacimovic, V., & Lakhani, K. R. (2024). The crowdless future? Generative AI and creative problem- solving. Organization Science, 35(5), 1589-1607. How can you boost your creativity?  Develop your Creative self efficacy Creative self-efficacy (CSE) is the confidence individuals have in their ability to generate creative ideas and solutions. It reflects a person’s belief in their capacity to engage in innovative thinking and overcome obstacles in the creative process. People with high CSE are often more willing to take risks, explore novel approaches, and persist through challenges, which can be crucial for successful innovation and problem- solving. They are innovative problem-solvers with unique skills relevant for every stage of the innovation process (Idea generation, Idea ) They have a prosocial mindset that motivates them to help others.  Connect with creative people Employees are more likely to improve their creativity when they interact with people who have innovation relevant skills and access to novel information. Research shows that the average CSE of alters in an employee’s problem solving network is positively related to that employee’s innovation behavior, This relationship is strengthened when the alters have a less dense network. having network contacts with high levels of CSE also leads to an increase in ego’s personal CSE one year later in cases where the employee’s initial level of CSE was relatively low. Grosser, T. J., Venkataramani, V., & Labianca, G. J. (2017). An alter-centric perspective on employee innovation: The importance of alters’ creative self-efficacy and network structure. Journal of Applied Psychology, 102(9), 1360. Improve your networking behavior Your creativity depends in your position on your network and the density of your network. Different measures of centralities (position) From immediate connections  Degree centrality (indegree/ outdegree) From the entire graph  Betweenness centrality (Broker)  Closeness centrality Different centralities and different roles in the network Sarah Higher degree centralities Ajay Spencer The broker in the advice network (highest betweenn Mike Zack Kelly Mary John Andrea Develop your creativity through leisure activities How can managers harness human-AI collaboration for creativity while preserving key human idea-generating skills?  Algorithmic Management Diminishes Perceived Creativity in the Workplace (Schweitzer and De Cremer 2024) In a simulated organizational setting, managers perceived employees led by algorithms as less creative compared to those led by human managers. This perception of reduced creativity affected managers’ views on the value and potential of algorithm-led teams. As a result, managers allocated fewer budget resources to teams managed by algorithms than to those led by humans. Schweitzer, S., & De Cremer, D. (2024). When being managed by technology: does algorithmic management affect perceptions of workers’ creative capacities?. Academy of Management Discoveries, 10(3), 375-392.  Class activity: Reflect and discuss Work in pairs to identify concrete actions that you can implement as managers to foster Human-AI creativity in an organizational context. Improving creativity in the organizational context  AI integration in the Business process  Promote an organizational context that fosters creativity  Responsible use of AI Integration of AI in the Business processes Companies should prioritize thoughtful integration of AI into their processes, moving beyond a narrow focus on automation. Long-term success lies in leveraging AI for augmentation and collaborative ensembling rather than emphasizing task specialization and full automation. Companies should only focus on automation and task division for routine and repetitive tasks. Encourage Human AI collaboration rather than task specialization to preserve human creativity. The need to reduce redundant work (automate work without added value for employees).  Promote an organizational context that fosters creativity Encourage creativity by facilitating serendipitous meetings where people can connect, discuss, and share ideas. Promote a culture of experimentation. Set up incentives to encourage employees propose new ideas Promote an environment where employees can thrive.  Make also decisions related to the use of Gen AI Use public Gen AI vs. implement a Corporate Gen AI. Invest in employee training to develop AI literacy and expertise in Gen AI use. Encourage mentoring between employees proficient in using Gen AI and those who need additional support. Promote responsible use of AI in the organization.  Responsible AI “We define a responsible AI (RAI) program as the comprehensive set of policies, standards, roles, processes, tools, training, and other elements that are put in place across an organization to ensure that all AI systems are designed, developed, deployed, and used in a responsible manner. “(Kiron and Mills, 2023, MIT SMR) Responsible AI constitutes a set of principles that ensure ethical, transparent and accountable use of AI technologies consistent with user expectations, organizational values and societal laws and norms (Accenture 2021) Responsible AI Robustness Fairness Transparency Accountability and safety Data Laws and Human Societal Well- Governance regulation Oversight being Conclusion  Takeaways As generative AI reshapes creative work, its impact will depend on how individuals and their teams engage with and integrate AI into their processes. Human involvement is deemed essential, as unquestioning reliance on AI without human oversight could lead to adverse outcomes. Human needs to develop their creativity. Fostering human-AI creativity requires organizational decisions that involves  selecting the appropriate generative AI tools to be used and establishing guidelines for their effective use (Responsible AI use:AI augmentation should uphold the authenticity, originality, and ownership of human creative work. )  Companies should prioritize AI integration for augmentation and collaboration (ensembling) for creative work and focus on automation and task specialization for repetitive and routine tasks.  Creating an organizational environment that nurtures creativity and experimentation.  Cultivating an Organizational Culture that Champions Well-Being and Workplace Happiness  The power of Human intelligence Do not outsource your creativity to AI Work actively on the development of your own creativity The future of creativity lies in collaboration between Human and AI. The importance of the organizational context in the promotion of creativity AT HOME WORLDWIDE* * Chez vous, partout dans le monde