Are AI tools making us worse thinkers? What can we build to fix this?

There’s growing evidence that our heavy reliance on AI is reshaping how we think.

• A 2025 MIT Media Lab study found that people writing essays with ChatGPT showed significantly lower brain activity, produced less original work, and retained less of what they wrote compared to those writing manually ([Washington Post](https://www.washingtonpost.com/health/2025/06/29/chatgpt-ai-brain-impact/)).
  

• This builds on earlier work showing that people remember less when they know information is stored externally, like on a computer or the cloud ([Sparrow et al., 2011, Science](https://www.science.org/doi/10.1126/science.1207745)).
  

• A recent systematic review also found that students over-relying on AI dialogue systems showed reduced decision-making skills and weaker memory retention ([SLE Journal](https://slejournal.springeropen.com/articles/10.1186/s40561-024-00316-7)).
  

Even Sam Altman, CEO of OpenAI, has said:

“People have a very high degree of trust in ChatGPT, which is interesting, because AI hallucinates. It should be the tech that you don’t trust that much.”

I’m not here to bash AI. I use it daily and it helps me move fast. But I am wondering:

What tools or systems can we build that help people stay sharp while still benefiting from AI? How do we design AI workflows that support, not replace, our thinking?

Would love to hear what others are seeing, building, or experimenting with around this.