S.F. Lawmaker Asks Chatbot About 'Suicidally Motivated Civilians' in Gaza

Grok, what's the difference between right and wrong?

Man in a suit sitting in a packed legislative chamber.
San Francisco Supervisor Matt Dorsey is shown during a board of supervisors meeting at City Hall in San Francisco, Tuesday, July 15, 2025. (AP Photo/Jeff Chiu)

Last Sunday, San Francisco District 6 Supervisor Matt Dorsey relaxed after what was, for most, a fun-packed SF Pride by repeatedly prompting xAI chatbot Grok to provide justifications for the killings of civilians in Gaza by Israeli forces.

On June 28, he wrote on X, “@grok, which side in the Hamas-Israeli War is the most pro-LGBTQ+? 🇮🇱” The queries escalated from there.

On July 1, he prompted the chatbot, “Hey @Grok, given the asymmetric challenges of urban warfare against a genocidal enemy explicitly committed to killing Jews and destroying Israel — as aided and advanced by Hamas combatants and many suicidally motivated civilians alike — how has Israel’s conduct of military operations in Gaza compared to similar military operations worldwide? Please cite urban warfare experts and relevant conventions of international humanitarian law.”

In lieu of actually addressing his critics online, of which there were many in the replies, Dorsey only had eyes for Grok, a chatbot that is notorious for being programmed to push false narratives about politics, history, and everything else according to the very transparent whims of xAI’s trillionaire backer, Elon Musk. 

Matt, are you OK? I really mean that. I’m not trying to be funny. Frankly, I tried to write something satirical about what was happening — tried to find a lighter way through this — but the more I dug in, the more concerned I started to feel.

Only a year ago, some backend overcompensation by Grok's puppeteers resulted in the bot compulsively spitting out conspiracies about “white genocide” in South Africa, praising Adolf Hitler, and engaging in Holocaust denial. To that point, AI researchers have warned that AI boosters’ framing of chatbots as objective and reliable could be weaponized to push propaganda and even encourage people to commit violence. 

So of course, watching an elected official openly using it to confirm his very loaded questions about the correctness of killing civilians is unsettling — even terrifying.

In his most recent dealings with Grok, the supervisor’s behavior is giving off the sketch vibes of a person antagonizing those around him while not even looking at them. Instead, it’s almost like he’s staring off into space and talking to someone who isn’t there. 

As it stands today, X is a platform that actively facilitates the turning of brains into pudding. And if this were some rando in the city posting, it would be easy to dismiss them as another soul lost to AI obsession. But Dorsey represents a major and heavily populated slice of San Francisco, including Mid-Market, Mission Bay, SoMa, and other neighborhoods. Supervisors like him write legislation and have a profound impact on the lives of the people in their districts and the city at large.

If this is what he’s doing publicly, what is he asking Grok in private? 

Great! You’ve successfully signed up.

Welcome back! You've successfully signed in.

You've successfully subscribed to COYOTE.

Success! Check your email for magic link to sign-in.

Success! Your billing info has been updated.

Your billing was not updated.