Liu, C., Denrell J., Luukkonen J., Chater, N. (2026). Using AI Can Stifle Innovation, But It Doesn’t Have To. Harvard Business Review, 104(4), 40–43.

These excerpts explore a growing tension in how organizations use AI: the same tools that increase individual productivity may quietly reduce collective learning and innovation. The authors argue that when “good enough” answers become cheap and immediate, people have less incentive to do the independent exploration that generates new knowledge. Their answer isn’t less AI, but better-designed interaction with it, building strategic friction into workflows so people contribute their own hypotheses, context, data and judgment rather than simply consuming finished outputs. The distinction between AI “builders” and “free riders” is particularly useful: are we using AI to extend our thinking, or to avoid doing the thinking?

Note: emphasis in quotations is mine.

… individuals in organizations decide whether to explore on their own or reuse what others have found … what happens when ‘good enough’ answers become essentially free.

… as reuse rose, independent exploration would fall and teams would start to converge on the same few approaches. Put simply, productivity goes up, but innovation quietly flattens.

A few people were doing the exploring: the rest were waiting to benefit from it.

GenAI clearly changes the economics of learning. If a plausible strategy memo appears in minutes, fewer people will spend days doing the harder work that produces the understanding, such as talking to customers, triangulating data, and pressure testing assumptions. When ‘good enough’ is free, fewer people pay the cost to find what’s better.

Avoiding the productivity trap

… a fundamental shift in how we value effort.

absorptive capacity: the ability to evaluate, adapt, and improve ideas rather than just copying them.

When people must invest some effort to use what others found, they do more independent checking and tinkering, which both helps them learn from others and produces new knowledge for others to learn from.

To participate in sharing sessions, you had to first show evidence of an independent attempt.

Building absorptive capacity

Friction doesn’t reduce learning; it raises the quantity and often the quality of what is shared.

Unfortunately corporate AI is often designed to remove the need for understanding. Leaders optimize for adoption and convenience, unintentionally eroding their organization’s absorptive capacity.

Reconisder optimizing solely for ease of use and start designing ‘strategic friction’ back into their systems.

Introducing strategic friction

• Set a rule requiring documented independent efforts.
• Design the AI to ask for users’ input before generating a full report.
• Build gated interfaces that unlock assistance only after users deposit their own context … the interface would require a uploading a baseline hypothesis and tagging three key variables or constraints.

How to spot the builders

Do they accept outputs as finished work, or do they interrogate them, test assumptions, and add original observations?

’Centaurs’ retained human control over both dimensions, using AI only for targeted assistance. ‘Cyborgs’ let humans drive the ‘what’ but gave AI significant control over the ‘how’, engaging in continuous, critical dialogue. ‘Self automators’ ceded control of both dimensions to AI and experienced that the researchers call ‘no skilling,’ developing neither domain expertise nor AI fluency.

Behaviors that distinguish builders from free-riders: selective delegation … builders use AI drafting, brainstorming, and pattern matching and do the judgement heavy work themselves. They ask the AI to explain its reasoning, point out inconsistencies, and feed it their own data.

The direction of information flow tells the story: Builders push context into the AI; free-riders pull finished outputs from it.

AI can make organizations faster immediately. The challenge for leaders is to prevent that speed from eroding absorptive capacity. Calibrated strategic friction – designed into tools, workflows, and hiring – helps ensure that organizations don’t just produce answers but keep gaining insight.