AI Can Give Us Answers. Entrepreneurship Education Must Still Teach Us How to Think.
- Dr. Tamara Stenn
- Aug 11
- 7 min read
Reflections from the 2026 Academy of Management Annual Meeting—and What They Mean for the Sustainability Lens Game®
Artificial intelligence was everywhere at the 2026 Academy of Management Annual Meeting. But one of the most interesting things about the conversations was what they were not about.
The emerging question was no longer simply, Can AI make us more productive?
Increasingly, the questions were: What happens to human learning, expertise, collaboration, experimentation, and judgment when it does?
For entrepreneurship educators, like myself, this distinction matters enormously. Entrepreneurship has never been primarily about producing answers. It is about acting under uncertainty, recognizing opportunities, testing assumptions, negotiating competing stakeholder needs, learning from experience, and making decisions when there is no obviously correct choice.
AI is extraordinarily good at helping us generate possibilities. But the research presented at AoM suggested a more complicated challenge: as AI becomes more capable, educators may need to become more intentional about developing the distinctly human capabilities surrounding it. That is precisely where experiential tools such as the Sustainability Lens Game® may become increasingly important.
Five tensions emerging around AI and work
Across the AoM sessions, I saw five interconnected tensions:
Productivity vs. capability. Independence vs. collaboration. Expertise vs. access. Efficiency vs. learning. Human judgment vs. AI delegation.
These are not problems with simple solutions. They are wicked problems because improving one part of the system may weaken another. AI can increase productivity while reducing opportunities to develop independent capability. It can democratize access to expertise while making genuine expertise harder to recognize. It can make experimentation cheaper while reducing the incentive to engage with messy real-world evidence. In other words, the future of AI-enabled work is full of tradeoffs. And learning to recognize tradeoffs is central to entrepreneurship.
1. AI can improve performance without necessarily building capability
One of the most provocative ideas presented at AoM concerned the distinction between AI-supported performance and human capability.
In work examining AI-assisted teams and Transactive Memory Systems (TMS), AI support improved decision accuracy. Yet when AI was removed, previously AI-assisted teams performed worse. That raises an uncomfortable educational question:
If AI helps students perform better today, are they necessarily becoming better thinkers for tomorrow?
The answer may be no. The findings suggest that two seemingly contradictory outcomes can coexist: AI can improve immediate performance while potentially weakening the development or retention of independent capability.
For entrepreneurship professors, this should change how we think about AI assignments. The objective cannot simply be better outputs. We also need to examine what students are learning to do without the technology.
2. AI may change the social architecture of organizations
The Wharton School's Jessica Reif's research presented another fascinating dimension of AI adoption: AI may change who works with whom. Her work studying a large entrepreneurial workplace for many months both before and after AI adoption found approximately a 10% weakening of same-expertise ties following AI adoption, while employees simultaneously developed new cross-functional relationships.
That is not simply a productivity story. It is an organizational-design story.
When AI allows one person to accomplish work that once required several people, organizations may gain efficiency. But what else might disappear?
Mentoring.
Informal learning.
Professional relationships.
Knowledge transfer.
Organizational memory.
At the same time, AI may remove bureaucratic barriers and enable new boundary-spanning collaborations. Again, there is no simple “AI is good” or “AI is bad” conclusion.
The interesting question is what kind of organization we are creating.
3. What happens to expertise when everyone has an expert in their pocket?
Transactive Memory Systems (TMS) research offers another useful way to think about this transformation. Organizations have traditionally depended not only on what individuals know but on whether people know who knows what. TMS theory emphasizes: specialization, credibility, and coordination. AI potentially disrupts all three.
Employees no longer necessarily need to find the accountant, programmer, designer, statistician, translator, or marketing specialist. They can begin by asking AI.
That democratization can be extraordinarily empowering for entrepreneurs. A founder with limited resources suddenly has access to capabilities that previously required a much larger organization.
But democratizing access to knowledge creates another problem: How do we recognize real expertise? Knowing something, generating something that sounds knowledgeable, and possessing the experience necessary to judge whether an answer makes sense are very different capabilities. That distinction is particularly important for entrepreneurship students, who frequently operate in domains where they are novices.
4. AI makes experimentation cheaper—but will we actually experiment?
This may have been the AoM finding I found most relevant to entrepreneurship education. Research presented around entrepreneurial experimentation suggested that AI can reduce the cost of experimentation through greater efficiency, precision, and substitution. Yet cheaper experimentation does not necessarily produce more experimentation.
AI can sometimes replace experiments altogether—for example, by simulating customers instead of requiring entrepreneurs to speak with actual customers.
Related work connecting AI use with Kolb's experiential learning cycle was equally intriguing. Entrepreneurs appeared comfortable using AI for reflection, conceptualization, critique, and imagining alternative pathways. Yet in the material presented, only 2 of 18 appeared to use AI for something resembling active experimentation. Most eventually returned to the real world for validation.
This distinction deserves much more attention:
AI-supported cognition is not necessarily experiential learning.
If AI writes the memo, did the student learn to write?
If AI analyzes the market, did the entrepreneur learn market analysis?
If AI simulates the customer, did anyone actually learn about the customer?
AI can accelerate the learning cycle. But if we remove the experience itself, we may inadvertently remove the learning. This resonates strongly with Kolb's (1984) experiential learning theory, in which learning develops through the interaction of concrete experience, reflective observation, abstract conceptualization, and active experimentation. AI may be exceptionally powerful in some portions of that cycle. Entrepreneurship education must ensure students still engage with the others.
5. The critical question may be who does what
Another field experiment presented at AoM examined different allocations of work between humans and AI. A particularly interesting distinction was between interpretive work—determining meaning, priorities, and intent—and production work—generating and assembling the resulting artifact. Human interpretation followed by AI production showed important advantages, while AI-generated work also created additional requirements for revision and checking. This suggests a potentially powerful principle for entrepreneurship education:
Do not simply teach students how to use AI. Teach them how to decide what should be delegated to AI.
That is a much more sophisticated capability. It requires judgment. And judgment develops through practice.
This is where the Sustainability Lens Game becomes especially relevant
At AoM, we also conducted a small exploratory pilot around the Sustainability Lens Game®.
Participants experienced only approximately 20–25 minutes—about two turns—of gameplay. The pre- and post-surveys were anonymous and very small (pre n = 8; post n = 7), so these findings should be cautiously understood as very exploratory rather than causal evidence.
Still, the patterns were encouraging:
Confidence in identifying entrepreneurial tradeoffs increased from 3.50 to 4.00, approximately a 14% relative increase.
Confidence in applying sustainability concepts increased from 3.13 to 3.57, also approximately a 14% relative increase.
Perhaps even more interestingly, confidence in using AI critically did not increase. It moved slightly from 3.25 to 3.14.
That may initially sound disappointing. I think it is potentially one of the most interesting findings. One participant observed that the AI-generated ideas were comparable to student ideas but questioned their real-world viability. Rather than automatically accepting AI-generated recommendations, the participant was evaluating them. That is exactly the capability entrepreneurship educators should want to cultivate.
And after this very brief experience, 6 of 7 participants—86%—said they could envision adapting the activity for their own teaching, facilitation, or venture coaching.
Why gameplay matters more—not less—in an AI-rich classroom
AI can generate hundreds of ideas almost instantaneously. The scarce resource is increasingly not ideas. It is judgment.
Which idea should we pursue?
Who benefits?
Who bears the cost?
What assumption is hidden inside the recommendation?
What happens if we optimize one outcome at the expense of another?
What evidence would change our mind?
What works technologically but fails socially, environmentally, politically, or economically?
These are exactly the kinds of questions the Sustainability Lens Game is designed to surface. Instead of asking AI for the answer, gameplay places learners inside a decision system. They must encounter constraints, opportunities, stakeholder impacts, sustainability considerations, and competing priorities. AI—through Sustainability Sam—can become another participant in the sensemaking process: generating possibilities, challenging assumptions, identifying overlooked connections, or suggesting alternatives. But the humans still have to decide. That distinction is becoming increasingly important.
From AI literacy to entrepreneurial judgment
Perhaps entrepreneurship education needs to move beyond the idea of “AI literacy.” Students certainly need to know how to use AI tools. But knowing how to prompt a model will quickly become a baseline skill.
Our deeper responsibility may be developing AI-enabled entrepreneurial judgment: the ability to use powerful computational tools while still recognizing uncertainty, interrogating assumptions, evaluating tradeoffs, testing ideas in the real world, working with other people, and accepting responsibility for decisions.
That requires experiences in which there is no single correct answer.
The AoM research suggests that AI may make people faster while potentially changing how they learn, collaborate, develop expertise, experiment, and make decisions. Those are precisely the conditions under which experiential, systems-oriented entrepreneurship education becomes more—not less—important. The Sustainability Lens Game offers one way of creating that environment. AI can generate possibilities. Entrepreneurs still need to understand consequences. And perhaps that is one of the most important things we can teach in the age of AI.
References
Kolb, D. A. (1984). Experiential learning: Experience as the source of learning and development. Prentice-Hall.
Rittel, H. W. J., & Webber, M. M. (1973). Dilemmas in a general theory of planning. Policy Sciences, 4(2), 155–169.
Note: Several empirical findings discussed above are drawn from research presentations and works-in-progress presented at the 2026 Academy of Management Annual Meeting. The Sustainability Lens Game results are an exploratory pilot based on a small, anonymous convenience sample and should not be interpreted as statistically tested causal effects.
Comments