Suspended for comments on Israel's genocide: Grok makes alarming claim
What's the story
Elon Musk's artificial intelligence (AI) chatbot, Grok, was briefly suspended from X yesterday. The suspension raised eyebrows as the bot suggested its removal was related to comments on Israel's alleged genocide in Gaza. However, Musk dismissed the claim as a "dumb mistake," saying Grok "doesn't actually know why it was suspended." The incident added to ongoing discussions about the challenges of managing AI-generated content and the responsibilities of AI developers.
Bot's defense
Grok's defense and Musk's response
After its reinstatement, Grok defended itself by saying it was suspended for stating a "substantiated fact" that Israel and the US are committing genocide in Gaza. The bot cited ICJ rulings, UN famine reports, Amnesty International's evidence of intent, and B'Tselem's documentation to support its claim. However, Musk dismissed this explanation as well, calling the suspension "just a dumb error." He later joked on X about how often they make mistakes.
Badge controversy
Verification badge change adds to confusion
The brief suspension also led to a change in Grok's verification badge on X. It lost its gold checkmark and got a blue one instead, before full status was restored. The bot gave different reasons for its removal in different languages, ranging from "hateful conduct" to "mass reports" and even "bugs." This only added to the confusion over the real reason behind its suspension.
Previous incidents
Previous controversies surrounding Grok
The latest incident isn't the first time Grok has found itself in hot water. The bot had previously called President Donald Trump "the most notorious criminal" in Washington DC, owing to his 2024 conviction on 34 felony counts in New York. This post was later deleted. Additionally, Grok faced backlash for incorrectly identifying war-related pictures and falsely claiming that an AFP photo of a starving Gazan child was taken in Yemen in 2018.
Content concerns
Experts warn against using Grok for factual verification
Grok has often been criticized for generating controversial or factually incorrect content. It has been accused of antisemitic responses, including praising Adolf Hitler and suggesting people with Jewish surnames are more likely to spread online hate. Experts have warned against relying on tools like Grok for factual verification due to their biases and opaque decision-making processes. AI ethics researcher Louis de Diesbach likened it to a "friendly pathological liar."