Tech broligarch Elon Musk is planning to revamp his AI chatbot because the latest model keeps churning out answers he doesn’t like.

Grok, xAI’s flagship programme, has lately not only produced material that is incorrect but also inconvenient facts that undercut or otherwise outright contradict the Tesla CEO’s hard-held far-right beliefs.

He added there was “far too much garbage in any foundation model trained on uncorrected data”—an apparent jab against the perceived “wokeness” of Grok’s chief competitor, Open AI’s ChatGPT—and invited followers to post “divisive facts” to feed into his revamp efforts.

Musk’s proposed overhaul follows a series of flubs that’ve seen Grok either posting false statements that humiliate the world’s richest man, or true statements that he simply doesn’t like.

These include the programme mistakenly confirming Musk had posted that he’d stolen the wife of White House aide Stephen Miller, when he hadn’t, and suggesting right-wing violence is “more frequent and deadly” than left-wing violence in the United States.

For what it’s worth, the latter appears to be largely true. A 2020 report by the Center for Strategic and International studies found that right-wing groups were responsible for almost 60 percent of the 893 terror attacks and plots orchestrated in the United States, compared to just 25 percent carried out by left-wing extremists.

The SpaceX founder’s prospective revamp of Grok is likely to have a sizable impact on the quality of public discourse on X, which critics warn has become a breeding ground for misinformation and hate speech since Musk’s takeover back in October 2022.

“Straight out of 1984,” Gary Marcus, professor emeritus at New York University, specialising in AI research, posted in response to Musk. “You couldn’t get Grok to align with your own personal beliefs, so you are going to rewrite history to make it conform to your views.”