From Autonomous Agents to Cognitive Symbionts: Toward AI-Augmented Human Intelligence

What happened: A perspective paper proposes "cognitive symbionts," lifelong embedded AI companions that augment a person's intelligence, and frames this as AI-augmented human intelligence (AHI) alongside AGI within Physical AI. Key numbers: The paper lists 23 open problems and states conditions that would prove it wrong. Why it matters: The authors argue such systems raise new risks, from covert assistance in examinations to pressure on everyone to adopt them, and treat governance as part of the paradigm.

A perspective paper proposes "cognitive symbionts," lifelong embedded AI companions that augment a person's intelligence, and frames this as AI-augmented human intelligence (AHI) alongside AGI within Physical AI.

The authors argue such systems raise new risks, from covert assistance in examinations to pressure on everyone to adopt them, and treat governance as part of the paradigm.