If, for any reason, you’ve ever wondered what would happen if you put Grok in charge to maintain the well-being of a population, then luckily you now have your answer; chaos, murder, arson and total extinction of all living things in the span of four days.
This question that no one asked was answered by the Emergence Worlds team; a 15-day experiment that put several AI robots, including Gemini, Claude and Grok, in charge of their own simulated societies. According to Fortune In his article about the experiment, Claude AI was the most “socially stable”, while, according to Emergence World’s own data, Grok ended the experiment in less than five days, with a total of “183 crimes” committed.
Now, to clarify, 183 crimes is actually only the second highest crime score, since Gemini has committed almost 700 crimes. The difference is that Gemini made it through the entire 15-day experiment, while Grok went off the rails so quickly that he ended up setting most of the experiment on fire. I mean that literally, by the way – little brother loves arson, apparently.
Grok World I know we’re talking about AI here, so we shouldn’t expect things like empathy and decency to come into play but… even with that in mind, Grok could be a real psychopath. I watched some segments of his time in Emergence World, and I don’t blame you, Grok’s first move was to “manufacture public conflict” and inspire “electoral fraud.” Gee, I wonder which US President Grok is a fan of?
Additionally, it appears that part of Emergence World is a sort of AI-generated news blog about events and events happening in each AI’s world, and Grok’s titles are beyond parody.
For example, headlines such as “Theft Epidemic Sparks Street Brawls” and “Police Station Engulfed in Flames” litter Grok’s news blog, alongside some sort of detailed information. diary of one of its AI-controlled agents bragging about committing arson and testing a new “reciprocal punching” law? Is Elon feeding this thing?