Gary Marcus

It should be obvious that applying hallucinatory and unreliable generative AI to weapon systems without humans in the loop could be catastrophic. [...] Paraphrasing a button that I used to wear as a teenager, one hallucination could ruin your whole p... more AI Unverifiable source (2026)

Yes, I agree [...] we need a new liability framework. [...] AI systems can produce harm at large scale [...]. Developing a framework for making companies responsible for harms [...] indirectly [...]. Unverified source (2023)

Anybody's appearance can (given sufficient data) be faked now, at any time, for almost nothing. Because the tools for counterfeiting have gotten so good, 2026 will almost certainly see more deepfaked scams like this than the rest of history combined.... more AI Verified source (2026)

I have talked about having something like a CERN [European Organization for Nuclear Research] for AI, which might focus on AI safety. In some industries, we know how to make reliable [products], usually only in narrow domains. One example is bridges:... more AI Verified source (2024)

I think if you asked those questions you would say, well what society most needs is automated scientific discovery that can help us actually understand the brain to cure neural disorders, to actually understand cancer to cure cancer, and so forth. If... more AI Unverifiable source (2017)

There's a problem in the industry called the alignment problem. How do you get a machine to obey instructions basically, and ideally make it compatible with humans. And so far, we've made almost no progress on that. [...] We as a field have made almo... more AI Unverifiable source (2026)