The Double Language of the AI Race

Linkedin

Prof. Lior Zalmanson

In the same week that The New Yorker published a tough investigation into Sam Altman, someone threw a Molotov cocktail at his home. Soon after, Altman shared a personal post with a photo of his family. He wrote that fear of AI is justified, and asked people to lower the flames.

That response is deeply understandable. After an event like that, anyone would want the escalation to stop. And yet, it is hard to read this without noticing this glaring and deeper contradiction.

For years, AI leaders have warned the public that this technology could reshape society, disrupt labor markets, and introduce risks on an unprecedented scale. They raised the temperature themselves. So when the public responds with fear, anxiety, or even hysteria, it is a little strange to ask everyone to calm down.

As Casey Newton recently argued, the tension around AI did not emerge on its own. It was actively cultivated. And at the very same time, many of the companies sounding the loudest alarms are also resisting meaningful oversight and regulation. So the public is being asked to believe two things at once: that the danger is real, and that there is no real need to slow down or submit to outside control.

When tobacco companies faced mounting evidence about the harms of smoking, they denied, obscured, and delayed. The AI industry is doing almost the opposite. The danger is not being hidden; it is being narrated. Sometimes even dramatized. The message is: this technology is enormously powerful and potentially dangerous, which is exactly why it must remain in the right hands. And somehow, those “right hands” always seem to be our own.

This logic appears inside organizations too. Employees are told to adopt AI quickly, move faster, and become more productive. But in the same breath, they are also told to verify everything, catch every error, and remain fully accountable for every output. That is a managerial fiction. You cannot meaningfully delegate cognitive work while pretending that nothing has been delegated.

So it's not really about if the AI race can be stopped. Almost no one with power seems interested in stopping. It's more about can we stop speaking in two languages.

If companies want to compete for power, they should say so openly. If they want to talk about societal risk, they should accept real oversight. And if organizations want workers to use AI, shouldn't they take responsibility for errors, review time, and the professional and psychological costs (and not just for the productivity boosts).

And the rest of us should probably ask ourselves a harder question too: do our values really align with how easily we keep handing over more and more of our judgment, attention, and lives to machines?