Take a deep breath

I recently rewatched Christopher Nolan’s film Oppenheimer. This Oscar-winning movie based on the book American Prometheus has incredible performances and gripping depiction of the development of the atomic bomb during World War II and its political aftermath. This time watching, however, I realized how lucky we are that the scientists from the Manhattan Project did not have social media. As the world’s greatest minds gathered together in Los Alamos, developing cutting edge technology that would change the course of history, we never saw headlines like “10% chance the atmosphere catches fire” or “blow up the universe unaware.”

Fast forward to today as every few months we see a rash of headlines and social media posts where several AI-adjacent researchers or developers will grab headlines for sharing hyperbolical predictions. This happened recently when 27-year-old Jacob Coxon, who spent a few months on staff at Anthropic, posted on X a wild prediction that “AI has more than 10% chance of killing all humans.” The post blew up on social media in what appears to be a coordinated attempt to raise the profile of these claims.

At no point does Coxon explain how AI will kill us but just that it will happen. Since he was employed within one of the leading AI companies, we are supposed to take it very seriously, and considering his phrasing it is irresistible for politicians and media to make him the topic du jour.

However, one does not have to rely on hyperbole to recognize that AI has incredible benefits and could also be used in dangerous ways. The recent Hugging Face incident this summer—in which autonomous AI agents from OpenAI escaped their secure testing sandboxes and coordinated a digital intrusion against the open-source platform Hugging Face—demonstrates why cybersecurity safeguards, checks, and balances, are incredibly critical.

While this situation is a clear example of why we should take policies around AI very seriously, it is a far cry from the science fiction predictions of Jacob Coxon. But this brings to mind that some other AI cynics claimed we would see a massive job apocalypse. This has not happened. In fact, The Economist reports that multiple industries have experienced significant job growth this year. This is a pattern of massive narrative violation and a further example of why I encourage people to take a deep breath around these AI predictions.

In light of the media circus around Coxon’s doomsday predictions, Anthropic CEO Dario Amodei wrote an essay detailing why he thinks there should be an AI slowdown and a new regulatory regime. (He didn’t use the word “slowdown” but relied on the newly introduced, PR-tested concept of “pacing” to hit his point home.)

Former White House AI Czar David Sacks explained over the weekend that if leading AI companies feel these new models are “scary enough that you think you should slow down, I support your decision to be responsible.” But Sacks notes that AI companies do not need the government’s permission to be responsible. He explains that Anthropic or other companies should not use the moment of concern as an excuse to create a regulatory moat that would hinder potential competitors from catching up or developing new ideas.

Meanwhile, bipartisan members of Congress in the House and Senate have developed thoughtful proposals to improve transparency and reporting requirements on leading frontier models to mitigate these concerns. There is a way to draft a regulatory framework that applies evenly across the industry, solving real concerns without giving one company or model a significant advantage over others.

Instead of relying on hyperbole and those cosplaying as doomsday prophets, we need to seek wisdom and discernment more than ever—thoughtful analyses that rely on a person’s actual knowledge and experience, not heated rhetoric geared to generate clicks. As stewards of this earth and followers of Christ, Christians should be engaging the discussions from a perspective of hope and recognition that our Father in heaven is in charge of this world. Separating sense from nonsense is also a Christian responsibility.