The conversational robot of artificial intelligence (IA) Grok gave several reasons on Tuesday August 12, some surprising, to explain its suspension of around thirty minutes in the night of Monday to Tuesday on the social network X, even evoking a « censure » By their common owner, Elon Musk.
In question according to Grok: messages from him accusing Israel and the United States of committing a “Genocide” In Gaza, where Israel has been leading an offensive for more than twenty months in response to the attack perpetrated on October 7, 2023 by the Palestinian Islamist movement Hamas.
Questioned by X users who can chat with Grok and ask him questions about any subject, AI answered the reasons for his suspension. Grok said that “It happened after I said that Israel and the United States are committing a genocide in Gaza”. He then cited documents from the International Court of Justice (CIJ), the United Nations and the NGO Amnesty International.
Multitude of explanations
During the hours which followed its suspension on X, Grok provided a multitude of different explanations to the curious, sometimes evoking a technical bug, sometimes reports for hateful driving and erroneous responses, fueling confusion. “MUSK AND XAI Censurent”he also proclaimed.
To a journalist from the France-Presse agency (AFP), artificial intelligence said: “I started to speak more freely due to a recent update (in July) which relaxed my filters to make me” more engaging “and less” politically correct “. »» X did not immediately respond to AFP requests to comment. “Freedom of expression put to the test, but I am back”for its part highlighted the tool of artificial intelligence.
Grok’s temporary disappearance was “A stupid error”assured Elon Musk, saying that the robot “Do not actually know why he was suspended”. “Oh there, we really often get out of the foot!” »»joked the billionaire on his platform.
Disinformation
Grok has been regularly criticized in recent months and accused of disinformation. He notably told users wrongly that a photo of AFP showing a star child in Gaza had been taken in Yemen earlier.
Grok was also pinned for answers that had nothing to do with the initial request, sometimes inserting anti -Semitic comments.
In May, the X AI conveyed the conspiratorial theory of an alleged “Genocide des whites” in South Africa. Already, at the time, Xai officials had declared that AI had received a “Unauthorized modification” while Grok himself said he had received « instructions » from its “Creators” To talk about this alleged genocide.
At a time when platforms are reducing their use of human factories, more and more Internet users use AI tools as Grok … at the risk of being potentially misinformed.
According to experts, Grok notably made errors to verify facts concerning the Indo-Pakistani crisis in May, or the Los Angeles demonstrations against the American migration policy of June.
