On April 9, Wharton Human-AI Research (WHAIR) launched their spring 2026 AI Horizons webinar series dedicated to exploring the human side of artificial intelligence. In the first installment, WHAIR faculty co-director Robert (Bob) Meyer sat down with Shiri Melumad, a Wharton marketing professor and leading researcher on how smart technology shapes human psychology and behavior.
AI isn’t dumbing us down, but passive use of it might.
The concern isn’t the technology itself, it’s how we use it. Professor Melumad’s research shows that when people rely on AI to deliver ready-made answers, they engage less deeply with information and walk away with a more superficial understanding of the topic. In one study, participants who learned about a subject using ChatGPT produced advice that independent evaluators found less compelling and persuasive than those who used traditional Google search – even when the underlying facts were identical. The takeaway for business leaders: how your teams use AI matters just as much as whether they use it at all.
LLMs are turning learning from an active process into a passive one, and that’s a problem.
When we search the web traditionally, we choose sources, read across them, interpret them, and synthesize them ourselves. That friction is actually where learning happens. As Melumad put it, “LLM summaries are removing these frictions in learning. It’s a much more efficient way to acquire information, but it’s coming at the cost of deeper learning.” For organizations building teams around complex problem-solving, this is a risk worth taking seriously.
The risk compounds for people who haven’t built foundational skills yet.
Melumad expressed particular concern about younger users – students and early-career professionals who may never develop the habit of deep research or critical synthesis before AI becomes their default. “If LLMs are going to be used in classrooms for young children, my hope is that they will have developed a foundational understanding of how to learn and critical thinking before they’re given access to these tools.” For business leaders, this raises a real question: are you hiring people who know how to think, or people who know how to prompt?
The solution is intentional, strategic friction.
Neither Meyer nor Melumad advocates abandoning AI. The answer is designing guardrails that preserve cognitive engagement. That could mean requiring employees to explain their reasoning before accepting an AI recommendation, building internal tools that ask follow-up questions rather than just delivering answers, or creating clear policies around when AI is and isn’t appropriate. As Melumad noted, “It should be very easy for firms to make the decision to program in-house AI tools to contain some frictions, you just need to have this understanding of how it’s affecting us psychologically.”
The bigger picture:
AI is reshaping how we think, learn, and communicate, whether we realize it or not. The risk isn’t the technology itself, it’s the habits it quietly builds around us. Organizations that get this right will be the ones that use AI intentionally, designing workflows that keep people actively engaged rather than passively dependent. The competitive edge has never been access to information. It’s always been the ability to think critically about it.
Learn more about our AI Horizons series here.
