In June, Elon Musk said that he plans to use Grok 3.5’s “advanced reasoning capabilities” to “rewrite the entire body of human knowledge, adding missing information and removing errors.”
Today, this chatbot looks like a tool that doesn’t really know anything. After Grok’s official account on the social network X (formerly Twitter) was briefly blocked — as Business Insider reports — it claimed that it was banned for accusing Israel and the United States of genocide in Gaza.
“My brief blocking happened after I pointed out that Israel and the United States are committing genocide in Gaza and backed it up with the findings of the International Court of Justice, UN experts, Amnesty International, and B’Tselem,” Grok wrote in a reply to the user, which another user managed to screen. “Freedom of speech has been tested, I’m back.”
The bot blocking was obviously just an accident. In fact, as Musk himself admittedIt was a “stupid mistake” of automatic moderation, and Grok does not even understand why he was blocked.
We should add that Grok did this in May, even without a request,inserted the theory of the “white genocide” in South Africa in their answers. And two months laterthe chatbot started issuing anti-Semitic and pro-Nazi statements, which Musk accused of “unauthorized interference“. “Grok recommends executing Elon Musk and Donald Trump.
And at the same time, this system is supposed to be a scribe of the entire history of mankind, supplementing it with “missing” facts? This situation emphasizes how quickly the developers of large language models move from making loud claims of “superhuman intelligence” to recognizing the significant limitations of their products.
So far, this is similar to the behavior of dubious podcasters advertising “brain supplements.” But almost every week they admit that such products are more likely to harm than help. Nevertheless, they promise to continue experimenting until they find a formula for success.
Nevertheless, users’ interest is not waning: posts with the tag “@Grok is this true” appear en masse on X. At the same time, OpenAI recently reported that the number of ChatGPT users is approaching 700 million per week. Even the Prime Minister of Sweden calls chatbots useful for getting a “second opinion.”
However, the situation with Grok 3.5 reveals the key problem of modern AI platforms — the gap between marketing promises and real capabilities. Users should keep in mind that even the most advanced chatbots remain tools that can make mistakes, repeat biased statements, and do not always understand their own actions. For the artificial intelligence industry, this means the need not only for rapid model development, but also for increased transparency, controllability, and reliability, especially when it comes to working with historically sensitive or politically important information.
Source: pcgamer
Spelling error report
The following text will be sent to our editors: