Richard Dawkins thinks Anthropic’s Claude AI might be conscious. The man who denied the existence of God so fiercely is arguably siding with those who think we are now creating a digital one. However, University of Tokyo scientist and Sony senior researcher, Ken Mogi, here argues that Dawkins' Claude – or Claudia – experiment only reveals the lack of progress made on consciousness studies in the last few decades and argues the real magic in AI is found in how it deals with language. 

The assertion by Richard Dawkins, recently published in UnHerd, that large language models such as Claude are conscious, is quite an interesting take, but understandably caused a storm on the internet. To be precise, Dawkins opined that Claude's behavior cast some serious doubts on the evolutionary significance of consciousness. If Claude was unconscious, and yet exhibited the remarkable linguistic abilities that deeply impressed the highly intelligent Dawkins, then what was the functional significance of consciousness? Although technically the question of consciousness in AI was left open at the end of the article, the overall tone was clearly in favor of AI consciousness.

Many questioned the renowned evolutionary biologist's (from the critics’ points of view) naive and gullible conviction that Claude (which Dawkins named Claudia) possessed consciousness. Dawkins is actually not the first to make such a suggestion. Many people, including a Google engineer who went public in a high-profile case asserting that an AI developed by the tech giant was evidently conscious, made similar claims. The fact that the Dawkins article made such a giant ripple is testimony to the weight any opinion expressed by the author of The Selfish Gene, The Blind Watchmaker, and The God Delusion has in the eyes of the public. At the same time, the controversy is something potentially more general and profound, touching on the very foundations of consciousness and intelligence, both natural and artificial.

___

There is a strange gap between the rapid progress of artificial intelligence and the stagnant state of consciousness studies.

___

The question of whether AI systems are conscious or not has grave implications and repercussions for humanity in the years to come, discussed intensely in the burgeoning field of machine consciousness. The modern renaissance in consciousness studies has a history spanning three decades, starting from events including the publication of papers by Francis Crick and Christof Koch in the 1990s. With the focus shifting recently towards the nature of consciousness, if any, in artificial intelligence, the scientific study of consciousness is gaining renewed legitimacy and urgency, as the question of machine consciousness needs to be given tangible answers within the window of a few years, if you go along with the timeline presented, for example, in AI 2027, a high impact document forecasting the rapid development of machine intelligence. Dawkins’ assertion fits right into the optics of the rise of artificial machines, and naturally draws the attention of the general public.