Scientific American Examines Shifting Meaning of 'AGI' Amid Lab Claims

A recent article published by Scientific American examines the growing use of the term "artificial general intelligence," or AGI, by major AI companies who have suggested the milestone may be near. According to the publication, the term was once largely confined to technical and academic discussions among computer scientists and researchers studying machine cognition.

The article states that the meaning of AGI has "slipped" from its original technical usage, raising questions about how the term is now applied in public statements by AI labs. Scientific American notes that as companies increasingly invoke AGI in describing their goals or progress, the definition attached to the term may differ from how it was originally understood in scientific literature.

The piece does not identify a single, universally accepted definition of AGI, reflecting a broader debate within the AI research community over what capabilities would need to be demonstrated for a system to be considered generally intelligent in a way comparable to humans.

Scientific American's analysis suggests that the gap between technical and public usage of the term could complicate efforts to evaluate claims made by AI developers about the imminence of AGI, since observers and companies may not be using the same benchmarks or criteria when discussing the concept.

The article was published as part of Scientific American's ongoing coverage of developments in artificial intelligence and does not name specific companies or provide detailed technical benchmarks in the excerpt reviewed.