Several writers have expressed their surprise at the sudden change in attitude among AI technologists. Allysia Finley observed, “They promised us the future, and now they want us terrified of it.” 

One example she uses is Bill Gates. Back in 2007, he said, “I think short of the transporter, most things you see in science fiction are, in the next decade, the kinds of things you’ll see.” The same Bill Gates has now been warning that AI could cause billions of deaths.

What is the origin of these fears? As far as I can tell, much of goes back to a fear about AI making paper clips. In his book from twelve years ago, Nick Bostrom imagined a superintelligent computer that was given the objective of maximizing the production of paper clips. One limiting factor would be human control of resources. Therefore, the logic of the computer would thereby conclude that eliminating humans would be an effective strategy.

Most people would naturally assume that programmers would include lots of safety protocols that would prevent this disaster. That is why Alysia Finley introduces you to Zack Korman, a self-described doomer who believes AI’s risks can be contained. He understands the risk and challenges but also argues that we need common sense cybersecurity.

One example has been the story of how OpenAI agents were able to break out of their testing “sandbox” and hacked into the machine-learning platform Hugging Face. He argues that if OpenAI had conducted “basic cybersecurity” this would not have happened. He also reminds us that we don’t necessarily need new laws since we have “a whole area of law that will regulate them into not killing people.”

AI poses a challenge and that is why technologists need to implement safety protocols.