cd /news/artificial-intelligence/your-brain-on-ai · home topics artificial-intelligence article
[ARTICLE · art-110909] src=technologyreview.com ↗ pub= topic=artificial-intelligence verified=true sentiment=· neutral

Your brain on AI

A study by Pattie Maes and colleagues at the MIT Media Lab found that people who used AI chatbots to evaluate news headlines were initially 21% more accurate at detecting fake news, but after four weeks they became 15% worse at identifying fake news without AI, despite about a quarter feeling they had improved. Lead author Anku Rani, a PhD student in media arts and sciences, said users forget that large language models are just statistical models predicting the next token, while co-lead author Valdemar Danry noted that AIs that ask Socratic questions foster learning better than those that give direct answers.

read2 min views2 publishedAug 25, 2026
Your brain on AI
Image: MIT Tech Review AI

Much as GPS weakens our navigation skills, relying on chatbots can undermine our ability to detect fake news.

Many people find AI-based chatbots helpful in keeping up with news, but a study by Pattie Maes and her colleagues at the MIT Media Lab points to a big problem with this strategy.

Participants who evaluated paired news headlines and images over the course of four weeks were initially 21% percent more accurate at telling fake news from real when aided by a chatbot—but by week four, they’d become 15% worse at identifying fake news without AI than they were before (though roughly a quarter of them reported feeling better at it). The result reflects an “AI dependency paradox” that has been observed in other domains, including medicine.

“Users get excited about these ‘magical’ LLMs but forget that they’re just statistical models that predict the next ‘token’ in a sequence,” says Anku Rani, a PhD student in media arts and sciences and one of the lead authors of the study, along with fellow MAS PhD student Valdemar Danry, SM ’23.

The researchers also found, however, that certain AI styles were associated with stronger independent performance later on, even if they slowed people down at first. “AIs that ‘tell’ by providing direct answers are more likely to foster reliance, while those that ‘ask’ via Socratic questioning are better at engaging someone to actually learn how to discern the truth on their own,” says Danry. “But it’s very much a trade-off between speed and effort.”

Keep Reading

Most Popular

A fundamental flaw leaves LLMs strikingly vulnerable to attack

It makes it easy to trick them into doing things they shouldn’t, such as telling you how to sabotage an aircraft’s navigation system.

Anthropic found a hidden space where Claude puzzles over concepts

A new technique has let the company probe deeper than ever into the weird workings of an LLM.

Sperm donors need limits, says a European fertility group

Some donor-conceived people are finding hundreds of siblings. An international cap on donations could help prevent that.

AI is more likely than humans to form biases when hiring

AI doesn’t just learn stereotypes from its training. It can cook up new ones, too.

Stay connected

Get the latest updates from #

MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

── more in #artificial-intelligence 4 stories · sorted by recency
── more on @pattie maes 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/your-brain-on-ai] indexed:0 read:2min 2026-08-25 ·