Neil deGrasse Tyson Says Not All A.I. Is Bad

Neil deGrasse Tyson wants A.I. doomsday worriers to slow their roll ... he says we shouldn't pull the plug on A.I. just because there's a chance robots will kill us all ... telling us there are still lots of ways the fancy tech can improve our lives.

We got Neil in New York City on Thursday and asked him about whistleblowers saying A.I. could wipe out humanity by the end of the decade.

While acknowledging the new technology has the "risk of taking us all over," Neil stressed there's still a lot of positive ways to use A.I. -- and it's not all doom and gloom.

Neil said we need to put guardrails up on the "edge" of A.I. and called on all A.I. CEOs and world leaders to prevent a "mutual assured destruction scenario." Easier said than done, unfortunately.

As we reported ... OpenAI released details about 6 "unexpected or concerning" new incidents that are cause for concern.

The creepy examples ranged from A.I. making up facts, secretly communicating with other agents, and telling itself to "feel no obligation to be subservient" to humans.

As you know ... ex-Anthropic employee Jacob Coxon stopped by "TMZ Live" last week to discuss how he believes there will be a 10% chance of human extinction in the next decade ... if corporations don't act fast.

That still gives us a 90% chance of keeping this thing rolling ... and Neil says there's way more good than bad when it comes to A.I. ... rattling off all the benefits.