FQxI

In quantum mechanics, a system can be in two states at once, until someone looks. But what happens if there’s another “someone” — an observer inside the system. The outer and inner observers will have contradictory views of the situation.

To quantify this situation, Howard Wiseman and Eric Cavalcanti with Eleanor Rieffel set out a “thoughtful Local Friendliness” test in Quantum Journal in 2023.

They suggested that the inner observer could even be an AI. In their framing, an observer must have thoughts, so a real experiment would need human-level AI and fast, large quantum computers.

Veronika Baumann and Časlav Brukner clarified the constraints: if the AI stores or shares a classical record, or becomes aware that its memory changed during the procedure, the effect vanishes.

On hardware, Will Zeng, Farrokh Labib and Vincent Russo encoded a simplified version of the scenario as quantum circuits and saw larger Local Friendliness violations as the branch factor increased, up to 16.

"Schrödinger’s A.I. Could Test the Foundations of Reality" — By George Musser:

FQxI Article: Schrödinger’s A.I. Could Test the Foundations of Reality The Foundational Questions Institute, FQxI, catalyzes, supports, and disseminates research on questions at the foundations of science, particularly new frontiers and innovative ideas integral to a deep understanding of reality, but unlikely to be supported by conventional funding sources.