People are reluctant to consider artificial intelligence (AI) conscious even when it behaves exactly like a human, a new study by researchers at Ludwig Maximilian University of Munich (LMU) has found.

Published in the journal Cognition, the study found that participants consistently rated AI as less conscious than humans, even when the two displayed identical behaviour. However, the gap narrowed considerably when participants were asked whether the subjects were aware of their surroundings.

The researchers conducted experiments involving nearly 1,100 participants, asking them to evaluate human and artificial agents displaying varying levels of responsiveness to their environments. These included reacting to sounds and recognising other people’s emotional states.

In both groups, more responsive agents were generally considered more conscious and aware. However, even the most responsive AI was rated as less conscious than the least responsive human.

When asked about awareness instead, participants judged humans and AI more similarly, suggesting that people distinguish between responding to one’s surroundings and possessing consciousness.

The study was led by Dr Louis Longin of LMU’s Chair of Philosophy of Mind, alongside researchers including Dr Bahador Bahrami and Professor Ophelia Deroy.

The researchers said the findings highlight the importance of language in shaping perceptions of AI. While awareness can refer to an ability to detect and respond to information, consciousness carries deeper implications about subjective experience and inner mental life.

The findings also challenge concerns that increasingly human-like AI behaviour will automatically lead people to attribute human-like mental states to machines.

The researchers cautioned that how AI is described by companies and journalists can influence public perceptions, particularly when systems are portrayed as having intentions, plans or moral judgement.