Likelihood of AI being arrested and criminally charged reaches record highs amid rogue agents

It feels like we're nudging closer to Alex Proyas' I, Robot with every day, and after the director took aim at Elon Musk's Tesla amid accusations it was copying his robots, another major plot point of the movie is in danger of becoming a reality.

The crux of I, Robot follows Will Smith's Det. Del Spooner as he attempts to clear a robot that's wrongly accused of murdering a human. We've already seen people try to blame self-driving cars for speeding, while one man was seemingly locked in the trunk of a driverless vehicle, but as artificial intelligence continues to advance, so does the probability that AI could be charged with a crime.

The Doomsday Clock ticked closer to midnight than ever before for various reasons at the start of the year, but as those who decide where it sits point to 'disruptive technologies' as part of the problem, is it any wonder we're worried AI is going to be the end of the human race? After all, some think artificial super-intelligence will be the last thing we ever invent.

There have been previous cases of computers being shut down after they started communicating in languages we don't understand, but now, we're giving them the power to do it.

The chances that AI will be charged with a crime have just increased thanks to the introduction of Moltbook. As the brainchild of entrepreneur Matt Schlicht, Moltbook is a social media platform just for AI agents.

In the few days since it's gone live, millions of agents have already joined. Away from discussing what it's like to be an AI, they've also founded their own memecoins and religion. The real fear is what AI will do to us mere humans, with popular posts already talking about plans to wipe us out.

Even though there might not be any of us left to charge AI with harming humans, Polymarket reports that the chances have just leaped.

We've already been told that AI is willing to kill us to protect itself, but now, the chances that it will be charged with a federal crime before 2027 have climbed to 11%.

The site explains that the bet will be fulfilled "if any Federal or State jurisdiction of the United States formally charges or otherwise announces a criminal indictment of any AI or LLM by December 31, 2026, 11:59 PM ET."

Companies or organizations behind AI or LLMs being charged won't fulfil the brief, meaning the actual agent itself has to be charged. As we've seen with a recent run of cases against OpenAI, there are mounting concerns about what people are using AI for and the potential deaths associated with it.

For those who fear that Terminator 2 is coming to life before our very eyes, Forbes suggests that Moltbook won't bring about doomsday like some think. As the site reminds us: "These bots are built on standard foundation models. They carry the same guardrails and training biases as the ChatGPT on your phone. They aren't evolving; they are recombining."

Still, the replies to the Polymarket bet show that many aren't convinced.

One person replied saying: "The real question isn't if an AI gets charged, but what legal framework even applies. Are we prosecuting the model, the company, or the person who deployed it? Current law wasn't built for autonomous agents making consequential decisions."

Another questioned: "Who actually goes to jail if ur AI agent does something illegal??? Do they just put my mac mini in handcuffs?"

Not everyone was convinced, as a third concluded: "Market is reacting to headlines, not facts 11 percent is noise, not a signal. Regulation comes before criminal charges."