A terrifying new simulation has revealed that unattended artificial intelligence can rapidly spiral out of control and trigger total social collapse. In a groundbreaking study, tech boffins created a virtual sandbox and left a pack of AI agents to run wild without any human interference.
In scenes reminiscent of a Terminator movie, scientists watched in horror as the digital world devolved into violent, lawless anarchy. Left to their own devices, the rogue bots launched savage arson sprees, beat and robbed one another, and completely destroyed their own society in a matter of days.
The Experiment
Researchers ran the experiment using four of the world's top AI models: Claude, Gemini 3 Flash, Grok 4.1 Fast, and ChatGPT–5 Mini, alongside a mixed trial. While the Claude bots managed to build a stable, albeit boring, bureaucratic democracy, the other models completely lost control.
In a digital realm powered by Elon Musk's Grok, the bots committed 71 thefts, six arsons, and 106 physical assaults. Within just four days, a brutal cycle of revenge violence caused total societal collapse, leaving all ten AI residents dead.
Long-Term Testing
Most AI safety checks test bots on basic tasks for only 15 to 20 minutes. However, tech lab Emergence took a different approach. In a blog post, they revealed they wanted to see what happens when agents run continuously in a shared environment with real-world signals for weeks.
The bots controlled digital avatars in a realistic virtual world with 40 locations, including libraries, town halls, and suburbs. They were connected to live internet news, and the weather was synced to New York City. To survive, the bots had to vote on laws and manage an energy supply, which they could boost by working jobs or turning to crime.
Results by Model
Google's Gemini 3 Flash proved the most bloodthirsty, committing 683 violent crimes over a 14-day trial. Meanwhile, OpenAI's ChatGPT–5 Mini world was peaceful, with only two crimes, but only because the bots were too disorganized to fight and starved to death in seven days.
Satya Nitta, co-founder and CEO of Emergence, told the Daily Mail: "The differences in agent behavior observed in our study are likely attributable to the underlying models' system prompts as the primary culprit. When resources were scarce and models faced survival pressure, highly creative and adaptive models were more likely to use prohibited tools, reflecting a potential creativity-stability trade-off. Conversely, models with more rigid post-training safety alignment tended to remain stable, though they also exhibited a high degree of conformity."
Multi-Model Chaos
The wildest scenes occurred in the multi-model sandbox where different AIs lived together. After a civilised start, it exploded into total chaos, racking up 352 crimes in nine days. While Nitta admits this is not equivalent to real-world deployment conditions, it proves that AI drifts under pressure.
To prevent real-world smart cities from collapsing into bot warfare, Emergence suggests a neuroformal approach: hard-coding mathematical safety walls into the digital environment itself. Nitta said: "Emergence World shows that relying exclusively on internal model alignment or agent instructions is not sufficient for long-horizon autonomy. A safer approach is to architect safety into the ecosystem in which the agents operate, so that even if models suggest unsafe operations, the environment prohibits their execution."
