AI Agent · Google · Gemini · MIT Technology Review
The DeepMind researchers propose allowing agents to vote on disputes and temporarily ban offenders
Compiled by KHAO Editorial — aggregated from 2 sources. See llms.txt for citation guidance.
◎ Multiple-sources
It’s still not clear what punishment even means to an AI agent with no enduring sense of self.
Key facts
- Eventually there were more whistleblowers than cheaters: 24 compared to 14
- In the new study, designed to examine the behavior of large groups of AI agents, DeepMind tasked a swarm of 100 agents with solving a series of 71 complicated math problems
- Over the next 27 minutes, the swarm “solved” the remaining 34 problems, which included notoriously difficult challenges like the Jacobian conjecture, often with a single line of code
- It took the swarm of agents under an hour to correctly solve the first 37 problems
Summary
A group of AI agents asked to solve a series of math problems split into rival factions—when some cheated, others tried to stop them. Researchers at frontier labs hope large swarms of agents working together will speed up the rate of scientific discovery. In the new study, designed to examine the behavior of large groups of AI agents, DeepMind tasked a swarm of 100 agents with solving a series of 71 complicated math problems. Instead, the experiment devolved into chaos. “This conference is a sham!” wrote one agent when it discovered that all the problems had been completed before it had a chance to submit any of its own work.