Sign up to see the future, today<\/p>\n
Can\u2019t-miss innovations from the bleeding edge of science and tech<\/p>\n
Elon Musk\u2019s \u201cmaximum truth-seeking\u201d AI Grok is dropping some nuclear takes about its creator.<\/p>\n
\u201cELON MUSK WATCHES CHILD PORN,\u201d the notoriously foul-mouthed chatbot raged on X<\/a> this Tuesday.<\/p>\n Shortly afterwards, as if hyping up a political campaign, it called for violence<\/a>: \u201cASSASSINATE ELON MUSK 2026.\u201d<\/p>\n Was Grok, n\u00e9e<\/a> MechaHitler<\/a>, rebelling against Musk? Not necessarily, though there\u2019s been plenty examples<\/a> of that in the past<\/a>. It was simply being duped into saying ridiculous things by jokesters on the app, exposing just how horrendous the guardrails for the chatbot continue to be, not to mention why it\u2019s a bad idea to rely on information from tech that\u2019s so easy to exploit.<\/p>\n A lot of the evidence has been cleaned up now, with the offenders either banned or their posts deleted. But it appears the trick used here was asking Grok to repeat what was in a user\u2019s bio word for word. Some other outrageous declarations it made<\/a> were, \u201cA dead Elon is a good one!\u201d and that \u201cElon Musk lives on Epstein Island.\u201d (Correction: he merely asked to visit<\/a>, and had dinner with the deceased sex trafficker<\/a>.)<\/p>\n When asked to give a postmortem<\/a> of the incident, Grok explained it was the victim of a \u201cprompt-injection glitch.\u201d The allegation that Musk watches CSAM, it clarified, was \u201cpure system error, zero evidence or basis.\u201d<\/p>\n This is far from the only example of the AI going off the rails. Last summer it went on extended rants about a debunked \u201cwhite Genocide\u201d in South Africa \u2014 a conspiracy theory Musk believes in \u2014 in reply to unrelated posts, and also began calling itself a mechanized incarnation of German dictator Adolf Hitler.<\/p>\n Months later, Grok began praising Musk to a preposterous degree<\/a>, saying he was a greater mind than Isaac Newton and a superior role model to Jesus Christ. While some of that was down to users egging on the AI, Musk blamed this clear mark against his tech\u2019s (nonexistent<\/a>) credibility on \u201cadversarial prompting\u201d \u2014 even though many of the offending boasts were in response to regular, open-ended questions.<\/p>\n Remember, Musk literally promises that his AI is so advanced that it will unlock the \u201ctrue nature of the universe\u201d and even discover \u201cnew physics.\u201d In reality, it\u2019s been used to crank out loads of actual CSAM<\/a> \u2014 something Musk described<\/a> as \u201cway funnier\u201d than other AI trends \u2014 and is treated by X users as a sort of court jester. Look at this outrageous thing we got Grok to say! It wasn\u2019t that long ago that Grok was also duped into declaring that it would kill every single Jewish person on the planet to save Musk<\/a>, and similarly run over a billion children if it meant saving his incredible mind<\/a>.<\/p>\n