DeepMind Proposes Symbiotic AI as Singularity Alternative

Three Google-affiliated researchers want to retire the image of AI as a single brain in a box getting smarter by itself. In an essay for the Deepmind Institute, Benjamin Bratton, Blaise Agüera y Arcas, and James Manyika argue that artificial general intelligence won't arrive as one machine crossing a threshold alone. It will come from networks: agents, people, and institutions working together, trading tasks, correcting each other, adapting over time. Some current AI systems already work this way, splitting problems across multiple models that coordinate like a team rather than one system doing everything.

The authors call this arrangement "Artificial Symbiotic Intelligence," and they frame it as a direct alternative to the singularity, the long-standing idea of a lone superintelligence recursively improving itself until it outpaces human control. Their claim is narrower but still a departure from how most people talk about AGI: intelligence, they argue, is something that happens between people and machines, not inside any one of them. That reframes the job facing researchers, who would need to govern and coordinate sprawling systems of human and machine agents instead of building a single isolated mind.