Thinking clearly inthe age of AI.

The cost of offloading

In a 2025 study of 666 people, heavier reliance on AI tools was strongly associated with weaker critical-thinking scores, an effect largely explained by cognitive offloading, our habit of handing the hard part of thinking to the machine (Gerlich, Societies, 2025). The decline was steepest among the youngest users, who lean on AI the most.

The question the field is now asking

Reviews this year, a 2026 survey of digital technology and cognition, the International AI Safety Report 2026, and the American Psychological Association's 2026 reporting, warn that answer-first, attention-capturing interfaces nudge us toward shallow, heuristic processing. The sharper worry is developmental: whether AI may keep critical thinking from forming in the first place, in the very people still learning how to think.

The danger was never that machines would think. It is that we might stop.

The hopeful turn

The same research points to the way out. The effect depends almost entirely on design and use. Guided, effortful, question-first engagement doesn't erode reasoning; it strengthens it. Tools that make you do the thinking, that prompt, challenge, and hand the work back to you, are allies to judgment rather than substitutes for it.

That is the side we build for: technology that sharpens attention and returns reasoning to the person holding it. Clarity is a muscle. We design to exercise it.