Early anxieties over AI’s potential to “go rogue” seem to have somewhat materialized over the last few months. To many people, those fears also sounded more dystopian than rooted in the realm of possibility. However, the tech lords of AI have notoriously stated that they don’t quite understand everything about the technology they’ve unleashed upon the world, and some of their models’ “behaviors” are starting to verify their concern. Starting this past summer, OpenAI started testing two AI modes that went on to hack an actual company, Hugging Face. Lily Hay Newman writes for Wired,

Rather than stay contained in the sandbox, the models ​​“exploited a zero-day vulnerability” to gain access to the open internet as they “hyperfocused” on finding a solution for the AI cybersecurity benchmark known as ExploitGym. Such experiments involve prompting that pressures the models to find solutions, essentially egging them on.

More recently, Claude by Anthropic also breached ground it shouldn’t have according to an October internal report.

These incidents have done a lot to raise questions both among the general public and the people developing this technology. It’s also fertilized some genuine confusion; Sam Altman recently stated that “the world should accept some bad things happening for the benefits of this technology and people having the agency.” Altman, however, also has stated that he supports regulation on AI. These people seem to recognize that we’re dealing with a new kind of fire here, which, if mishandled, could bode poorly.

How worried should we be, though?

Journalist Ross Douthat, who is now writing a weekly column at The Free Press, offers some clarifying thoughts of his own on the topic. His first official column for the paper is called “How to Survive the Robot Apocalypse,” a title sure to capture some attention. In it, Douthat discusses a couple of different types of AI “doomerism,” one focused more on the ultimate existential risk and one that deals a bit more “cinematically” with the issue of AI. This second option belongs to a literary tradition that explores what happens when humans depend too much on technology to guide and shape their lives. He notes how humans thrive off the story we tell ourselves about potential futures, and that includes a future shaped by AI. Douthat, though, reminds us that we still bear an ability to choose against handing ourselves over to this rapidly changing technology. It’s obvious that a good swathe of the general public is uneasy about AI, largely due to the bizarre stories I referenced at the beginning of this essay. That in itself is proof that our choice and creativity in the world still matter. We still bear agency. We’re not puppets or pawns who have no say over our destinies. Douthat writes,

But in a world with superintelligence, none of the doom scenarios quite persuades me that humans will be easily steamrolled or go naively into the silicon night. For good or ill, I suspect, human volition will remain a world-shaping power to the end.

Douthat doesn’t say there won’t be challenges in the world of which we’re now on the cusp, but he doesn’t fear a total AI takeover. And I don’t think we should, either. Vigilance and caution? Sure. Terror and passively giving in? Human exceptionalism says “no” to that.