Inferring Consciousness Out There
Marcello Massimini, Giulio Tononi
Oxford Scholarship Online July 19, 2018 DOI: 10.1093/oso/9780198728443.003.0008 (opens in new tab)
Study at a glance
AI-extracted from the abstract| Characteristics | Theoretical or philosophical paper Peer reviewed |
|---|---|
| Key points | Argues that behavioral complexity and brain size are unreliable criteria for consciousness in animals, and that large artificial systems like feed-forward deep learning networks may not be conscious despite their performance. Proposes that future empirical measurements of information integration may help infer consciousness in non-human entities. |
Abstract
This chapter describes a preliminary exploration in search of criteria for consciousness in the biological and physical world beyond the human skull, including dolphins, octopuses, parrots, bees, and computers. It argues that assessing the complexity of behavior and measuring the size of the brain may not provide a reliable estimate in animals. Likewise, it explains why some artificial systems, such as such as feed-forward deep learning networks, that are composed by many elements and that perform incredible feats, may not be conscious. Finally, it suggests that in the future precise empirical measurements of information integration may offer a valid tool to infer on the capacity of consciousness in non-human entities.