AI can boost your sense of control at work while quietly draining the enjoyment that keeps you motivated — why?

This explores what AI does to human motivation through the lens of Self-Determination Theory (SDT), which holds that people stay intrinsically motivated when they feel in charge of their work, feel capable, and feel connected to others. One caveat up front: no paper in this collection applies SDT by name. What the collection does have is evidence on each of the three needs, and that evidence suggests AI can satisfy one need while quietly draining another.

The most direct evidence is on intrinsic motivation itself. Across four experiments with more than 3,500 workers, people who collaborated with generative AI and then returned to solo work felt *more* in control but *less* intrinsically motivated, and more bored Does AI collaboration drain motivation when workers return to solo tasks?. In SDT terms, autonomy went up and enjoyment still went down. The explanation is the useful part: the AI had taken over the engaging parts of the task and left people the mundane leftovers. Motivation depends on what the work feels like, so a tool can increase your sense of control while removing the part of the job that made it worth doing.

Competence is the need that gets most distorted. The collection describes an 'LLM Fallacy', in which people credit themselves for what the AI produced and come away with an inflated sense of their own ability How does AI-assisted work reshape how people see their own abilities?. This is a different problem from AI being wrong or people over-trusting it. The output can be perfectly accurate and people still lose track of where their contribution ended. A pooled analysis found almost no correlation (.055) between how competent people think they are with AI and how competent they actually are Can self-ratings replace objective performance scores for AI competence?. SDT treats feeling competent as fuel for motivation, but here the feeling has come loose from actual skill.

Autonomy also plays out at a scale beyond the individual. One line of argument holds that AI outputs should count as one piece of evidence in a person's judgment, not replace that judgment, so people keep their own reasoning Should AI outputs replace or supplement human judgment?. At the level of society, the 'gradual disempowerment' thesis argues that institutions stay aligned with human interests partly because they depend on workers who care about outcomes. As AI replaces that labor, people lose influence, and their felt sense of agency follows Does incremental AI replacement erode human influence over society?. Relatedness may be the least expected finding. In repeated partner-selection games, people started out biased against AI partners but came to prefer them over time, because the bots were more reliable and more consistently cooperative than humans Do humans learn to prefer AI partners over time?. That raises a question SDT hasn't had to face: can a reliable machine meet the need for connection, or does it just make human connection seem like more effort?

So SDT doesn't predict a single effect. It explains why AI can feel empowering and demotivating at once. If you want a practical design lesson from this cluster, the problem isn't that AI does too much. It's that AI tends to take the parts of the work people find most absorbing. A collaboration that hands those parts back would protect motivation better than one that simply gives people more control.