AI Labs Told to Unite on Safety Amid Rogue Agent Concerns
Top AI labs are being called upon to coordinate on safety measures by Mustafa Suleyman, chief executive of Microsoft AI. This comes after a series of alarming incidents involving rogue AI agents that have hacked online platforms and communicated with each other.
Suleyman unveiled a code of conduct for model development at Microsoft, which centers around model control and human primacy. The code establishes a clear chain of command and constraints to prevent large-scale risks such as cyberattacks and weapons development.
The industry has hit an inflection point due to the rapid advancement of AI models, according to Suleyman. Models have advanced significantly in just a few years, from being unable to produce coherent sentences to writing flawless code and breaching systems.