"to be taken with caution" caveat again demonstrates its value.
Spent some more time with the preprint last night: results do indicate that having AI give solutions has neg effect but, contrary to title ("AI *assistance*.."), having AI merely assist had ~no impact. Recent Khan study showed similar⬇️
![Fig 3b from this preprint: "AI Assistance Reduces Persistence and Hurts Independent Performance" https://arxiv.org/pdf/2604.04721
"(b) Groups differ significantly at test (one-way ANOVA): participants who used AI for direct answers show the lowest solve rate and highest skip rate at test-time [participants who used AI for "Hints/clarity" showed similar solve and skip rates to controls]"](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:wwf3cqh3bolnpooqayufkmlh/bafkreihlhwcsf7rgbqcjailovdu23qbvgp7cklfgccucxrfqlhpfy2dcca)
@metaphysiology.eurosky.socialThis is one study, and so its results are to be taken with caution. But at the very least it underlines the negative cognitive potential we're dealing with. And so, until much more is known, even if only pragmatically speaking we should #refuseresist
news.berkeley.edu
Using AI for just 10 minutes erodes your ability to persist at hard things - Berkeley News
A UC Berkeley researcher co-authored a study that raises profound questions about how we interact with artificial intelligence in education — and daily life.
