Google DeepMind co-founders Demis Hassabis and Shane Legg, alongside Google executive James Manyika, launched the DeepMind Institute on Wednesday to study artificial general intelligence governance.
The institute will function as a global research platform to study risks, including the potential "loss of control" as AI approaches human-level cognitive capabilities, the trio said in a statement.
"We’re already seeing the implications of increasingly capable AI for cybersecurity and biorisks, and there is the potential for loss of control in future self-improving systems," the executives said. "This is a critical moment to ensure we build AGI safely and its benefits to society far outweigh any risks."
The three executives will direct the institute, which will bring together researchers from Google DeepMind, Google and the wider global research community.
According to the statement, the DeepMind Institute will also examine how AGI affects what it means to be human, how systems and the communities of agents they form should be governed and which institutions and policies society may need to adapt or reimagine.
The directors said technologists should not provide all the answers. "The responsibility of shaping its reality collectively belongs to society as a whole, including the arts and humanities, and governments," they wrote.
However, the executives acknowledged that current AI systems can still fail at basic tasks and lack the consistency and creativity needed to meet the threshold for full AGI.
"While AI can still sometimes fail at basic tasks," the executives wrote, "we expect those gaps to be closed soon."
The launch comes as concerns over advanced AI risks gain wider attention. On Thursday, King Charles warned AI executives of "existential dangers" if the technology falls into the wrong hands and urged safeguards before it becomes "too late."