Listen: Are we training AI to think too much like humans?
- James Evans, a University of Chicago sociologist and data scientist, believes that current AI models are being trained to think too much like humans.
- This narrow approach is holding back scientific exploration and limiting our ability to discover new breakthroughs.
- Evans advocates for building AIs that think “differently” from humans, which he calls “cognitive aliens”.
- The goal of these cognitive aliens would be to unlock hidden discoveries by thinking outside the box and exploring unconventional ideas.
- Building such AI systems would require a fundamental shift in how we design and train them, moving away from human-centric approaches and towards more innovative and creative solutions.
James Evans believes we’re training AI to think too much like humans—and it’s holding science back.
In this episode of the Big Brains podcast, Evans, a University of Chicago sociologist and data scientist, shares how our current models risk narrowing scientific exploration rather than expanding it.
He also digs into why he’s pushing for AIs that think differently from us—what he calls “cognitive aliens.”
Could these “alien minds” help us unlock hidden breakthroughs? And what would it take to build them?
Read the transcript of this episode.
Source: University of Chicago
The post Listen: Are we training AI to think too much like humans? appeared first on Futurity.
Q. What does James Evans believe is a limitation of current AI models?
A. He believes that our current models think too much like humans, which may be holding science back.
Q. What profession does James Evans hold at the University of Chicago?
A. He is both a sociologist and a data scientist.
Q. According to James Evans, what are the risks of training AI to think too much like humans?
A. The risk is that it narrows scientific exploration rather than expanding it.
Q. What term does James Evans use to describe AIs that think differently from humans?
A. He calls them “cognitive aliens”.
Q. Could cognitive aliens help us unlock hidden breakthroughs in science?
A. Yes, according to James Evans.
Q. What would it take to build cognitive aliens?
A. The text does not provide specific details on what it would take, but rather invites the reader to explore this idea further.
Q. Why is James Evans pushing for AIs that think differently from humans?
A. He wants to encourage scientific exploration and discovery.
Q. Where can we find a transcript of an episode featuring James Evans discussing his views on AI?
A. The transcript is available on the Futurity website.
Q. What institution does James Evans work at?
A. He works at the University of Chicago.
Q. Is James Evans’ view that current AI models are too similar to humans a criticism or a suggestion for improvement?
A. It appears to be both, as he suggests that AIs should think differently from humans while also acknowledging the limitations of current models.