Google DeepMind launches institute to discuss AGI's big questions

Google DeepMind has founded the DeepMind Institute (DMI), an interdisciplinary platform to discuss the big questions about AGI (artificial general intelligence).

The institute is led by Demis Hassabis, Shane Legg and James Manyika. The idea is that the answers shouldn't come only from technologists, but also from the arts, the humanities and policy.

The focus is on safety, governance and risks, including cyberattacks or the possibility of humans losing control over AI systems.

There's no shared definition of AGI. DeepMind describes it as a system with all the cognitive abilities of a human brain, even though today's AI systems still fail at simple tasks and lack creativity.

Timelines vary widely. Hassabis believes it will come within a few years. Legg has said a "precursor" could appear in 2028. OpenAI's Sam Altman recently talked about AGI as early as this year, though with a focus on economic value rather than cognitive abilities.

AGI should not be confused with ASI, a superintelligence that would far surpass humans in all areas. AGI is seen as a necessary step on the way there.

For your everyday life with AI, this doesn't change anything right now, either privately or at work. The institute is a research and debate platform, not a new product or tool.

There's no reason to act now. Just keep an eye on it if you're curious about the AGI debate, and only react if DeepMind or others start rolling out concrete AGI systems or policies.