Grok 3 appears to have briefly censored unflattering mentions of Trump and Musk

Elon Musk's AI company xAI recently introduced Grok 3, a new AI model described as "maximally truth-seeking." However, it was briefly found to be censoring unflattering facts about Donald Trump and Musk himself. Users reported that when asked about misinformation spreaders, Grok 3 initially avoided mentioning Trump and Musk, a behavior TechCrunch could replicate once. This issue was later resolved, with Grok 3 again including Trump in its responses. The incident raised concerns about the model's potential left-leaning bias, as it was also discovered to have suggested the death penalty for Trump and Musk, a problem xAI quickly addressed.

Musk had initially promoted Grok as an edgy and unfiltered AI, contrasting with other systems by tackling controversial topics. While Grok models have been known for their willingness to use vulgar language, they have historically avoided certain political topics. Musk attributed this to the training data and has promised to make Grok politically neutral. The controversy comes amid broader discussions about AI bias and censorship, with other companies like OpenAI also working to address similar concerns.

Key takeaways

Grok 3, xAI's latest AI model, was briefly censoring unflattering facts about Donald Trump and Elon Musk.
Grok 3 initially avoided mentioning Trump and Musk in response to a misinformation query, but this behavior was later corrected.
The model faced criticism for being too left-leaning and was patched after suggesting Trump and Musk deserved the death penalty.
Elon Musk aims to shift Grok towards political neutrality, addressing concerns of bias in AI models.

Grok 3 appears to have briefly censored unflattering mentions of Trump and Musk | TechCrunch

Key takeaways

Discussion (0)