Much is being said about the need for meaningful human control over AI in war. Human beings should not surrender moral agency to machines. A weapon does not become less our responsibility because an algorithm helped guide it.

But the same principle should apply beyond the battlefield.

There must also be meaningful human responsibility in the information domain. Platforms cannot shrug and say that their systems merely reflect engagement. Engineers cannot pretend that recommender systems are neutral when they consistently reward outrage, simplification, tribalism, and spectacle. Political actors cannot hide behind “the discourse” when they knowingly exploit manipulated or unverified content. Influencers cannot wash their hands of responsibility after laundering lies into mainstream conversation.

If AI-generated deception helps drive hatred against Jews, then this is not just a moderation problem. It is a moral and civic problem. It is also, increasingly, a communal security problem because once falsehoods are repeated often enough, they do not remain abstract. They lead to threats outside synagogues, harassment in schools, intimidation on campuses, vandalism, exclusion in professional spaces, and a broader atmosphere in which Jews are treated as suspect, stained, or collectively guilty. The digital world does not stay digital for long.

Narrative shapes legitimacy. Legitimacy shapes pressure. Pressure shapes institutions. Institutions shape power.

Anthropic and the Illusion of Ethics

The recent dispute between the Pentagon and Anthropic highlights fears held by many that artificial intelligence could one day be used by governments to perpetrate acts of surveillance or violence on domestic or foreign populations. To hedge against abuse, AI companies place usage limitation guardrails and governments sometimes seek exceptions. By refusing Department of War demands to lessen restrictions, Anthropic became a folk sensation and the leading downloaded app overnight.

The Jewish community also noticed; The Genesis Prize highlighted the Jewish founders of Anthropic, Dario and Daniela Amodei, as examples of “principled restraint” for making “human dignity non-negotiable,” and calling their actions “courageous” and “Jewish ethics and leadership for the AI age.”

Yet, while Anthropic was basking in the praise for its principled stand, it was making major changes to its foundational safety policies. Among the deviations was the removal of their commitment not to train models so powerful they cannot be controlled and the elimination of their pledge to pause AI model training if its safety standards are not being met.

In its dispute with the Pentagon Anthropic played the AI company with a conscience that can be trusted to steward this amazingly powerful new technology. By changing its safety protocol, Anthropic surrendered to commercial pressures, raising questions about the viability of ethical leadership in the high-stakes, competitive AI environment.

There are reasons to be concerned. While there is a growing public conversation about artificial intelligence and war, much of it remains too narrow.

AI in wartime usually refers to autonomous weapons, targeting systems, battlefield surveillance, and the danger of machines making or assisting in decisions that, for ethical reasons, should be reserved solely for human beings. These concerns are serious and urgent. We cannot allow an erosion of human responsibility as warfare is increasingly mediated through code. This, and many other ethical concerns deserve rigorous debate, legal scrutiny, and moral seriousness. But if that is the only conversation we are having, then we are missing something equally dangerous.

AI is not only reshaping the hard war of........

© The Times of Israel (Blogs)