Google DeepMind, together with partners, has announced a $10 million funding call dedicated to multi-agent AI safety research. The programme is intended to support external work on the unique risks that can arise when multiple AI systems interact with each other or with people.
The announcement frames multi-agent systems as an increasingly important area, since these setups introduce coordination and emergent behaviour that single-agent research may not cover. By opening the call to outside researchers, DeepMind is signalling that safety in this area needs a broader community effort.
No specific research topics, deadlines, or application details are included in the source. The only details given are the funding amountcars and the fact that DeepMind and partners are behind the call.