Will AI kill all humans? Anthropic employees freak out!

Is it time for the killer robot nightmare? Man oh man, a lot of very smart people certainly seem to think so!

So, a bunch of former, and in some cases current, staff at AI company Anthropic are alleging that they’re really scared that artificial intelligence is going to go rogue and kill everyone — and yet some of them are charging boldly ahead, still working on the project.

Jacob Coxon, who resigned from Anthropic, wrote the following on X:

“The people building AI earnestly believe that it could kill us all by the end of the decade. This is not a marketing stunt. If anything, many executives and senior researchers will couch their phrasing in the press to sound sensible – but I hear the same people express fear privately. No other human activity poses this level of danger.”

And in response, current Anthropic staffer Evan Hubinger wrote :

“Jacob is correct here—we really do earnestly believe AI could kill all humans! I personally think it is >10% within the next decade. I believe Anthropic is trying its best, but we do not yet have a plan to solve alignment for superintelligence and are not clearly on track to.”

First, why do you believe there’s more than a 1 in 10 chance that AI will kill all humans — all 8 billion of us? What would make it want to do that, and what is the mechanism by which it would do that?

Next, what are the odds it is able to save the lives of hundreds of millions of humans, by curing cancer etc.?

And then, what are the odds of a foreign power designing an AI that could kill all of us, immaterial of what we do? And is our best option for stopping a foreign adversarial AI to come up with our own, better AI?

It’s worth keeping in mind that higher intelligence generally correlates with less violence and more humane behavior. We have invented scary technologies before, including nuclear technology, and this has not led to the extinction of our species. On the contrary, since the advent of nuclear technology, there has been a period of relative peace. War is less common today than it was 100 years ago and before that.

Reacting to the news, Florida Gov. Ron DeSantis had this to say :

“While some AI hysterics is aimed towards investors in advance of upcoming IPOs, that there is even a possibility that these technologies would elude human control is alarming. Technology should enhance the human experience; it should never supplant it!”

And I agree with that. I agree, based on the concerns we are hearing, we need sober analysis and common-sense intervention to ensure that superintelligence has proper guardrails in place.

What we shouldn’t do is blindly panic. We should do as Ted Cruz suggests , and make sure the American continues to lead on this front.

So Republicans are expressing the appropriate level of concern. You know who isn’t expressing the appropriate level of concern? Democrats. Bernie Sanders wants a ban on AI outright! By which I mean a ban on American-led AI — China is going to keep on keeping on.

And Michigan Senate candidate Abdul El Sayed said the same thing. In reaction to the latest news, he told reporters that it’s time to hit the emergency stop button and shut it all down. This is what Democrats will do if they regain power: let China get an edge on this technology, and make the U.S. play catchup. That is a very, very dangerous strategy: It’s one that has a much higher risk of dooming us.

Bottom line: Republicans understand the assignment. Democrats, however, are ready to shut it all down.

Robby Soave is co-host of The Hill’s commentary show “Rising” and a senior editor for Reason Magazine. This column is an edited transcription of his daily commentary.