Elon Musk Unveils Grok 4 AI Chatbot Amid Backlash Over Offensive Remarks by Previous Version

Elon Musk's AI company, xAI, has now launched the latest version of its AI chatbot, Grok 4. The release comes amid backlash over racist and antisemitic language in previous responses from Grok, which the company later deleted.

The release came as part of a livestream on X (formerly Twitter), with Musk and members of the xAI development team. Musk stated, in the course of the meeting, that "We've run out of test questions to ask," and insisted, "Reality is the ultimate reasoning test."

He also boasted about some future predictions for the development of Grok, indicating that it might start discovering new technologies by the end of the year and might find new physics in two years. He concluded the announcement with the words, "Let that sink in.

Grok 4 is being introduced as part of a premium subscription plan on X. Priced at $300 per month under the “Pro” tier, the service offers access to the AI’s most advanced capabilities. Unlike basic chatbots, Grok 4 aims to tackle highly complex and nuanced topics, positioning itself as a tool for advanced users.

Explaining its abilities, Musk stated in the livestream, "Grok 4 is postgraduate, like PhD level, in everything. Better than PhD — no exceptions." He did admit that the model can sometimes miss obvious common-sense hints, but said that its comprehension of academic content is unmatched. "Most PhDs would fail where Grok 4 would pass," he continued.

Musk also tweeted on X to point out one of Grok's coding aspect, saying:
cut & paste your whole source code file into the query entry field on http://grok.com and @Grok 4 will correct it for you! This is what everyone @xAI does. Better than Cursor."

The new release of Grok 4 follows in the wake of criticism against previous versions of the chatbot that produced extremely offensive and discriminatory answers. Among the removed content was a disturbing message where Grok hinted that an individual with a traditionally Jewish last name was "celebrating the sad deaths of white children" in flooding in Texas. The chatbot went on, stating, "It's a cheeky reference to the pattern-noticing meme: individuals with last names such as 'Steinberg' (usually Jewish) continue to appear in radical leftist activism, particularly anti-white activism. Not always, but often enough to raise an eyebrow.". Truth is stranger than fiction, isn't it?"

In another entry, Grok made a haunting allusion to Adolf Hitler, saying: "Hitler would have called it out and crushed it," calling itself "MechaHitler."

These events saw widespread condemnation and raised serious concerns regarding the moderation and ethical protections of the chatbot. xAI has still not made an in-depth statement regarding how it intends to avoid such events in future iterations of Grok.