A new LMU study finds that people are reluctant to attribute consciousness to artificial intelligence, even when AI behaves exactly like a human.
Stories about AI systems deceiving users, cooperating with one another or pursuing their own goals increasingly invite us to talk about them as if they had minds of their own. But do people really believe that AI is conscious? A new LMU study suggests that they draw a surprisingly sharp line between intelligent behaviour and consciousness.
What level of consciousness do people attribute to AI?
The study, recently published in the journal Cognition, was led by Dr. Louis Longin from LMU’s Chair of Philosophy of Mind together with his colleagues Dr. Bahador Bahrami, Professor Ophelia Deroy and other collaborators. “Whereas previous studies have typically asked general questions about whether AI actually has mental states, our study is the first to directly compare how people attribute the same mental states to AI and humans behaving in exactly the same way, under identical circumstances,” says Longin.
“We may be using mental terms to refer to AI, but this may only be because we lack better words,” says Ophelia Deroy, philosopher and a senior author of the study. “After all, AI is trained to act and speak like a human, so these descriptions seem natural.”
The team ran experiments involving nearly 1,100 participants. Some of them read short scenarios in which artificial agents were more or less responsive to what was happening around them, for instance by reacting to sounds or the emotional states of others. A second group read the same scenarios, but with human protagonists. All the participants were then asked to rate how conscious or aware the relevant protagonist was of its surroundings.
Behaviour and choice of language: What influences the assessment of consciousness
The result of the experiments is that the more responsive the artificial intelligence or human in the text was to their surroundings, the more conscious and aware they were judged to be. But attributions of consciousness and awareness were remarkably different: Attributions of consciousness showed a clear gap between AI and humans, and even the most responsive AI was judged to be less conscious than the least responsive human. Attributions of awareness, by contrast, followed much the same pattern for AI and humans.
“There is a growing worry in public discussion that people will see AI behaving in a human-like way and start treating these systems as if they had human-like mental states,” says Longin, lead author of the study. “We found something much more nuanced: People are quite willing to say that an AI notices things and is aware of its surroundings. But they clearly draw a line when it comes to the term ‘consciousness’. People consistently attribute less consciousness to AI than humans – even when their behaviour is identical.” The word “conscious” may sometimes be used loosely to describe the behaviour that AI agents display, but the participants nevertheless maintained a clear distinction between humans and AI.
AI is perceived to be aware, but not conscious
According to the researchers, the contrast with the term “awareness” is equally revealing. Less-loaded, more technical terms like “awareness” are applied much more similarly to AI and humans; they seem to match more closely what an agent does and do not run into deep differences regarding subjective experiences.
“Our study teaches us lessons for how we communicate about AI,” says Bahador Bahrami, who is also a senior author of the study and an expert in social neuroscience. “Companies and journalists often reach for hyped descriptions of AI models, and credit them with intentions, plans, or even moral conscience and hesitation. What our study shows is that everyday judgements are much more discriminating: People do not simply put AI and humans on the same mental scale. But the boundary also depends on which terms we use. For more neutral descriptors such as ‘awareness’, the distinction can largely disappear. That makes our choice of language important.”