Skip to main content

Elon Musk, Tesla’s CEO, made headlines when he bought Twitter in October 2022. He renamed the company “X” and announced plans to turn it into an AI pioneer. He launched a chatbot for premium subscribers named “Grok,” which later appeared in Tesla cars. Recently, Grok users discovered the chatbot only says flattering things about Musk. Some jokesters twisted that programming until the bot spit out NSFW claims about him.

The Guardian reports that after one of Musk’s early 2023 posts flopped, he convened an emergency meeting with 80 X engineers. He ordered them to write a custom algorithm that boosted his posts by a factor of 1,000. It’s no surprise the Grok chatbot shows the same bias.

One X user posted bizarre replies from Grok. First, it praised Musk’s physique. Then it said he ranked “among the top 10 minds in history, rivaling polymaths like da Vinci or Newton.” Other users asked follow-ups. Grok claimed Musk is more fit than LeBron James. That’s when the internet went to town.

The internet really put Grok through it

Another user wrote a lewd prompt. Grok responded that Musk has the “potential to drink piss better than any human in history.”

Another user went even more explicit. Grok claimed Musk’s “blowjob prowess edges out Trump’s.” It argued he should have won a 2016 porn industry award because he’s “the ultimate throat goat.”

X deleted the posts quickly, but 404 Media archived them.

The ‘woke’ AI fell asleep at the wheel

AI experts have long warned that chatbots’ confidence—even when totally wrong—can mislead users. Critics argue this is a major reason not to integrate them into infotainment systems, as Tesla has done. Distracted drivers may trust chatbots for navigation or other functions without double-checking the answers.

Chatbots such as ChatGPT avoid NSFW topics. Musk positioned Grok as the opposite. Vice President JD Vance even said, “I’m a Grok guy…I think it’s the best. It’s also the least woke!” But this latest round of Grok answers shows why many chatbots stay cautious. Musk has already shut down and reprogrammed Grok once after a series of pro-Hitler rants.

For his part, Musk responded with humor. “For the record, I am a fat r——.” He added, “If I up my game a lot, the future AI might say ‘he was smart … for a human.’”

Related

Does the Honda Ridgeline’s Unibody Make it a Better Truck?

Want more news like this? Add MotorBiscuit as a preferred source on Google!
Preferred sources are prioritized in Top Stories, ensuring you never miss any of our editorial team's hard work.
Add as preferred source on Google