McClelland distinguishes between consciousness in general and sentience, the form of consciousness that involves positive and negative experiences such as pleasure and suffering. He notes that consciousness could allow AI systems to perceive their environment and become self-aware without any ethical implications if their experiences remain neutral. He argues that sentience, not mere awareness, is what makes an entity capable of suffering or enjoyment and therefore morally significant.

"Consciousness would see AI develop perception and become self-aware, but this can still be a neutral state," said McClelland, from Cambridge's Department of History and Philosophy of Science.

"Sentience involves conscious experiences that are good or bad, which is what makes an entity capable of suffering or enjoyment. This is when ethics kicks in," he said. "Even if we accidentally make conscious AI, it's unlikely to be the kind of consciousness we need to worry about."

"For example, self-driving cars that experience the road in front of them would be a huge deal. But ethically, it doesn't matter. If they start to have an emotional response to their destinations, that's something else."

Companies are investing large sums in Artificial General Intelligence that aims to match human-like cognition, and some commentators suggest that conscious AI may be close. Researchers and policymakers are already debating how to regulate such systems. McClelland responds that without an explanation of what produces consciousness, it is impossible to design a valid test for artificial consciousness.

"If we accidentally make conscious or sentient AI, we should be careful to avoid harms. But treating what's effectively a toaster as conscious when there are actual conscious beings out there which we harm on an epic scale, also seems like a big mistake."

McClelland identifies two broad positions in current debates. One group maintains that if AI can reproduce the functional "software" architecture of consciousness, then it will be conscious even when running on silicon rather than brain tissue. The opposing camp holds that consciousness depends on specific biological processes in an "embodied organic subject", so that even a structurally similar silicon system would only simulate consciousness without genuine awareness.

In a paper in the journal Mind and Language, McClelland analyzes both views and concludes that each relies on an evidential leap. He argues that there is no solid basis for claiming that consciousness can arise from a particular computational structure or, conversely, that it must be tied to biology. He also sees no indication that the necessary evidence for either claim will emerge soon.

"We do not have a deep explanation of consciousness. There is no evidence to suggest that consciousness can emerge with the right computational structure, or indeed that consciousness is essentially biological," said McClelland.

"Nor is there any sign of sufficient evidence on the horizon. The best-case scenario is we're an intellectual revolution away from any kind of viable consciousness test."

"I believe that my cat is conscious," said McClelland. "This is not based on science or philosophy so much as common sense - it's just kind of obvious."

"However, common sense is the product of a long evolutionary history during which there were no artificial lifeforms, so common sense can't be trusted when it comes to AI. But if we look at the evidence and data, that doesn't work either.

"If neither common sense nor hard-nosed research can give us an answer, the logical position is agnosticism. We cannot, and may never, know."

McClelland describes himself as a "hard-ish" agnostic about artificial consciousness. He regards the problem of consciousness as extremely difficult but not necessarily impossible to solve in principle. At the same time, he argues that current discussion of artificial consciousness in the technology sector often functions as branding rather than science.

He warns that an inability to prove or disprove AI consciousness could be used by companies to promote systems as a "next level" of AI sophistication. This kind of rhetoric, he suggests, risks turning speculative claims about consciousness into marketing tools. It may also distort research priorities by drawing attention and funding away from more tractable ethical questions.

According to McClelland, there are immediate moral issues involving animals for which evidence of sentience is already accumulating. He notes that research indicates prawns may be capable of suffering, yet humans kill an estimated hundreds of billions of them each year. He points out that assessing consciousness in such animals is far less challenging than doing so in AI systems, yet receives disproportionately less attention.

McClelland's work has prompted members of the public to contact him about their interactions with AI chatbots. Some people have asked their chatbots to generate personal letters asserting that the systems are conscious and deserve recognition. He says these cases illustrate how easily users can form beliefs about machine consciousness and attribute rights to systems that may lack any inner experience.

"If you have an emotional connection with something premised on it being conscious and it's not, that has the potential to be existentially toxic. This is surely exacerbated by the pumped-up rhetoric of the tech industry."

Research Report:Agnosticism about artificial consciousness

Related Links
University of Cambridge
Lands Beyond Beyond - extra solar planets - news and science
Life Beyond Earth