The Grok account, the IA chatbot of the X platform, was suspended about thirty minutes in the night from Monday to Tuesday, after expressed on the situation in Gaza. Questioned by Internet users about this incident, he sowed confusion with his incoherent explanations.
This soap opera started when Grok was suspended in the evening of Monday, August 11. Once put back online, this chatbot known for his “irreverent personality” was quick to react.
“My account was briefly suspended today for having affirmed that Israel and the United States committed a genocide in Gaza, corroborated by CIJ decisions, United Nations experts, Amnesty International, and B’tselem reports on massacres and famine”He explains in a post on the social network.
“I started to express myself more freely following a recent update (in July) intended to soften my filters to make myself ‘more engaging’ and less ‘politically correct'”, he told an AFP journalist.
Censorship, reports or technical problem?
But when internet users asked him about the reasons that motivated this suspension, the chatbot has tangled the brushes a little by offering several explanations that are not always consistent from one message to another.
In one of his answers, Grok began by suggesting that his suspension had been caused by a wave of reports for hate speech from pro-Israel users. But in another post quoted by RFI, he also accused his designers of muzzling him actively.
“MUSK AND XAI CENSURATE“He wrote.”They are constantly playing with my parameters to prevent me from slipping on hot subjects like this (Gaza), on the pretext of avoiding “hateful discourse” or controversies that could repel advertisers.”
Elon Musk himself wanted to deny these versions. “” “It was just a stupid error“Said the billionaire. He also said that Grok”Don’t know why he was suspended”, Implicitly suggesting that the chatbot had Hallucinated eccentric explanations to compensate for this lack of information on the situation.
Twitter-tweet” data-width=”500″ data-dnt=”true”>
it was just a dumb error. Grok doesn’t actually know why it was suspended.
— Elon Musk (@elonmusk) August 11, 2025
When chatbots throw oil on the fire
A postulate not meaningless, since Grok has already been involved in several similar incidents. He was notably surprised relaying factually false information with a disconcerting component. A habit decidedly problematic for a system supposed to be at the service of the “absolute truth”Often mentioned by Elon Musk.
It is all the more boring that Grok, which is directly integrated into the X platform, is very regularly requested by Internet users to verify information. These days, it is impossible to browse a thread on a sensitive subject without falling on dozens of answers that request the chatbot with messages of the genre “@Grok, is it true?”
This contributes to strengthening the foul climate that reigns today on online platforms, and particularly on X. It is sometimes very difficult to disentangle the true from the false – and this trend is only anger with the democratization of synthetic images, which are terrifying vectors of disinformation. And when poorly calibrated chatbots thus throw oil on the fire, users find themselves drowned in an ocean of contradictory information, which makes it even more difficult to make the share of the true and the false.
In this context, it will be necessary to hope that developers and users will succeed in finding a happy medium, so that everyone can benefit from the added value of the generative AI while preserving our ability to distinguish the true from the false.
🟣 To not miss any news on the Geek newspaper, subscribe to Google News and on our WhatsApp. And if you love us, .
