ResearchDiscussion

Is AI Making You Stupid? The Science of Cognitive Offloading

Dave Asprey

The speaker discusses a MIT study on how AI tools like ChatGPT affect brain network activation during cognitive tasks. They argue that the impact depends heavily on how users engage with AI, suggesting that most common usage patterns are likely harmful to cognitive function.

Summary

The conversation opens with a direct question about how harmful ChatGPT is for the brain. The speaker responds that the answer is highly dependent on how individuals engage with the tool, expressing the view that most people's typical usage patterns are probably detrimental to brain health, while acknowledging that more evidence is still needed.

The speaker references a notable MIT study examining brain network activation in individuals performing cognitive tasks with and without the assistance of ChatGPT. This study gained significant attention on social media, though the speaker notes it had not yet been formally published at the time, existing only as a preprint. The speaker also points out that despite the viral attention, most people likely did not read the full study due to its considerable length — approximately 220 pages — describing it as a very intensive piece of research.

The transcript ends before the speaker fully explains the specific findings or reasoning behind their concerns, but the setup suggests they were about to elaborate on why passive or over-reliant use of AI tools may negatively impact cognitive engagement and brain development.

Key Insights

  • The speaker argues that whether ChatGPT is harmful to the brain depends entirely on how a person engages with it, implying that passive or over-reliant use is the problematic pattern.
  • The speaker claims that the way most people currently engage with ChatGPT is probably bad for the brain, though they acknowledge the scientific evidence is still emerging.
  • The speaker references a MIT study that examined brain network activation in individuals performing cognitive tasks with and without ChatGPT assistance, framing it as key evidence for their position.
  • The MIT study went viral on social media, but the speaker suggests most people did not actually read it, noting it is approximately 220 pages long and described as a very intense study.
  • The study had not yet been formally published at the time of discussion, existing only as a preprint, which the speaker acknowledges while still treating it as significant evidence.

Topics

Cognitive offloadingAI and brain healthMIT ChatGPT study

Transcript

[0:00] How bad is chat GPD for your brain? >> It entirely depends on how you engage with it. I think the way that most people engage with it, it's probably really bad for your brain. We're still waiting for like more evidence to come in, but I'll tell you exactly why I think that's the case. When, you know, they've done studies looking at brain network activation in in individuals who are doing some kind of cognitive task with or without the help of of chat GPD. So there was this big famous study done at MIT. I'm not sure if it's even been formally published yet, but it came out online as a preprint. But you know, social…

Full transcript available for MurmurCast members

Sign Up to Access

More from Dave Asprey

Get AI summaries like this delivered to your inbox daily

Get AI summaries delivered to your inbox

MurmurCast summarizes your YouTube channels, podcasts, and newsletters into one daily email digest.