The Real AI Threat Is Already Here

Even Luddites terrified by the destructive capabilities of AI would have a hard time crafting nightmare scenarios as frightening as the ones conjured up by the industry itself. On September 8, researcher Jacob Coxon announced on X that he was resigning from the AI giant Anthropic. Coxon explained that Anthropic and another tech giant, OpenAI, were not “acting responsibly” but rather “are racing straight to self-improving superintelligence and gambling with our lives…. The people building AI earnestly believe that it could kill us all by the end of the decade.”

Coxon’s views were endorsed by Anthropic researcher Evan Hubinger, who wrote on X: “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.”

Coxon and Hubinger are far from alone; a surprisingly large number of researchers in the field have offered scenarios in which AI poses an existential threat, some of which eerily call to mind movies like The Terminator: AI develops consciousness, decides that it wants to be free of its pesky biological progenitor, and then figures out some plan to wipe out humanity. The possible extermination schemes include hacking nuclear-weapons systems, developing dangerous viruses, or simply shutting down the cybernetic infrastructure that is now essential to our existence.

As AI has increased in political salience, a wide swath of the public has given credence to these fears. According to data scientist David Shor, “64% of voters think it’s either very or somewhat likely that AI could pose a threat to humanity’s survival.”

Not everyone is afraid. The venture capitalist David Sacks, an ally of Elon Musk and an AI adviser to Donald Trump, argued on X that “Anthropic is running a sophisticated regulatory capture strategy based on fear-mongering.” In other words, the big AI firms are trying to deploy the fear of their product to scare politicians into agreeing to regulations that will effectively allow existing firms to have permanent cartel power over the technology.

On the opposite end of the political spectrum, the radical writer Richard Seymour also cautioned that AI threat inflation is a form of corporate hype. Writing in Jacobin, Seymour astutely pointed out, “Tech fetishism, the ascription of human agentive powers to machinery, is always a political act. It obscures that the goals, the purposes embodied in any technology, are human purposes.”

Most pertinently, a political discourse dominated by speculation about the existential dangers of out-of-control AI helps us avoid thinking about the problems caused by AI that is fully obedient to its human masters. In truth, the killer robots are already here and helping to commit genocides—not because they went haywire, but because they followed human orders.

In 2024, the Israeli journalist and filmmaker Yuval Abraham reported in the Israeli-Palestinian publication +972 that the Israel Defense Forces were using an AI tool called Lavender in Gaza. According to Abraham, Lavender not only developed the kill lists used by the IDF in the early stages of Israel’s response to the October 7 attack, but it was also a key instrument in a strategy that made large civilian casualties inevitable. “The Israeli army,” he notes, “systematically attacked the targeted individuals while they were in their homes—usually at night while their whole families were present—rather than during the course of military activity.” As part of this strategy, “automated systems, including one called ‘Where’s Daddy?,’…were used specifically to track the targeted individuals and carry out bombings when they had entered their family’s residences.”

Abraham’s reporting in +972 forms part of the basis for NAZA, a new documentary he made with Rachel Szor that won a special jury prize at the Venice Film Festival. (He also won an Oscar last year for codirecting the 2024 documentary No Other Land.) NAZA will undoubtedly spark a broader discussion not just on the Gaza genocide but also on the future of war—and AI’s central role in that future.

Israel, of course, is not alone: Many other countries are harnessing AI to kill people. Both the United States and Iran, for instance, have reportedly deployed Anthropic’s AI models in their ongoing war. But the scale of the atrocities taking place in Gaza puts Israel at the forefront of this dreadful trend, and what’s happening in Gaza will almost certainly spread throughout the rest of the world.

As the investigative journalist Antony Loewenstein documented in his 2023 book The Palestine Laboratory, “the Israeli state has become the research…pioneer for the most technologically advanced methods of human slaughter.” This has allowed Israel to develop technologies that it can sell abroad, including to autocracies such as the United Arab Emirates and Bahrain. For any dictatorship, Israel serves not just as a willing vendor for the tools of social control but also as a model for successful repression—a blight unto the nations. The “tools of war” used in Gaza, Loewenstein notes, “were proudly displayed on social media for both a domestic and international audience as well as potential global buyers. Israel used [AI]-enabled weapons, striking non-military targets with unprecedented ferocity.”

Of course, the AI researchers’ fears can’t be completely dismissed as fearmongering. AI is clearly the most dangerous technology that humankind has developed since nuclear weapons, and our current lack of oversight for it is a travesty. Just as it would be absurd to allow the private ownership of atomic bombs, AI companies need to be not only regulated but nationalized, and we need an international nonproliferation infrastructure as well.

But Gaza reminds us that simply controlling the weapons of war isn’t enough: We also need to punish those who use these deadly tools to commit war crimes. Right now, the real threat from AI is not the destruction of human agency but rather the way it amplifies the worst tendencies of humankind.