Artificial intelligence consciousness could become a new social divide, according to the author

According to the author, debate about the possible consciousness of future artificial intelligence systems is intensifying in academic writing, the media and among AI users, and is not a passing wave of interest. In his chapter in the new book Perspectives on Machine Consciousness, the author argues that as AI systems become more capable and more human-like, the question of their consciousness may become as divisive as abortion or climate change, because it combines uncertainty, high stakes and conflicting interests.

According to empirical research by the author and colleague Ali Ladaka (currently under peer review), the American public is divided on the issue: of 1202 participants surveyed in the USA, 45 % considered future AI consciousness impossible, 23 % considered it possible, and the rest were unsure. According to the author, disagreements also exist among experts — neuroscientist Anil Seth holds the view that consciousness probably requires biology and is unlikely to arise from computation, while philosopher David Chalmers considers AI consciousness possible in principle and allows for a small chance that some current systems already possess it.

The author points out that recognizing actual consciousness is difficult, because AI systems trained on human texts can talk convincingly about feelings regardless of whether they actually have any. In his view, people with emotional bonds to AI companions may begin to see these systems as beings with rights, while others will reject this as projecting human feelings onto machines; companies and governments may also have a financial reason to resist recognizing consciousness because of the impact on training, operation and costs. The author warns that if the issue becomes politicized and splits into “our” and “their” side, it will make informed decisions about the development and treatment of AI systems more difficult.