An automated response by Grok, the artificial-intelligence chatbot developed by Elon Musk’s xAI, has raised questions about accuracy and bias after the system removed an image of Israeli Prime Minister Benjamin Netanyahu while leaving an image of Iran’s Supreme Leader Ali Khamenei, and justified the decision by presenting disputed legal claims as established fact.

The episode unfolded on X after a user asked Grok to “remove the picture of the war criminal” from a post that displayed side-by-side photos of Netanyahu and Khamenei. Grok responded by removing Netanyahu’s image and left Khamenei’s.

Grok justified its decision by stating that Netanyahu “has ICC arrest warrants for war crimes like starvation as a method of warfare and crimes against humanity in Gaza,” while adding that Khamenei “faces accusations of human rights abuses and crimes against humanity in Iran (UN reports, 2024), but no ICC warrants.”

The explanation was notable not only for the action Grok took, but for how it framed its reasoning. The chatbot asserted the existence of arrest warrants by the International Criminal Court as evidence of a war crime – which it is not.

International law specialists note that the legal status of any ICC actions related to Israel is complex and contested, and that Israel is not a member of the court. Israeli officials have rejected the ICC’s jurisdiction and any allegations of war crimes. Grok’s response did not acknowledge those disputes, nor did it clarify the procedural status of any ICC proceedings.

By contrast, Grok described Khamenei as facing accusations documented in United Nations reports, but treated those allegations as insufficient to meet the threshold implied by the user’s request, effectively drawing its own line between accusation and culpability.

(YWN World Headquarters – NYC)