Can AI Have a Conscience? The Illusion of Machine Morality
Nearly 1,100 participants in a recent study drew a surprisingly sharp line between intelligent behavior and consciousness when evaluating artificial intelligence, according to a study published in the journal Cognition and reported by Newswise.
LMU Experiments Test Public Perception of Artificial Intelligence
The study was led by Dr. Louis Longin from LMU’s Chair of Philosophy of Mind, alongside Dr. Bahador Bahrami, Professor Ophelia Deroy, and other collaborators. Researchers tested whether people genuinely attribute minds to AI systems, even as 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.
Participants read short scenarios describing artificial agents that reacted to their surroundings, including environmental sounds and the emotional states of others. A second group evaluated identical scenarios featuring human protagonists. Following the readings, all participants rated how conscious or aware the relevant protagonist was of its surroundings. The team ran experiments involving nearly 1,100 participants.
Behavior Drives Awareness Ratings But Fails to Grant Consciousness
When an artificial intelligence reacted dynamically to its environment, observers readily acknowledged that the system noticed things around it.
However, that willingness stopped abruptly at the threshold of subjective experience. Participants consistently attributed less consciousness to AI than humans, even when their behaviour was identical. Lead author Dr. Louis Longin noted that while public discussion often worries about people treating machines as if they possess human-like mental states, the experimental data revealed a much more nuanced psychological boundary. 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. Longin also noted, After all, AI is trained to act and speak like a human, so these descriptions seem natural.
Religious Scholars Help Anthropic Instill Morality into AI
Parallel discussions over machine capabilities have drawn religious scholars into discussions with technology developers. A recent gathering brought religious scholars into discussions with Anthropic to examine whether they could help instill morality into Artificial Intelligence. Elizabeth Dias wrote a lengthy piece about these meetings in the New York Times. Those deliberations intersected with ongoing concerns regarding who controls the foundational values encoded into machine learning models.
Critics point out that attempts to align automated systems with human values risk imposing the specific moral visions of a few onto broader populations. It is noted that while engineers express awe at the “beauty” they see inside artificial “neural networks,” fundamental questions remain regarding whether artificial architecture can ever replicate the nuanced, evolving nature of human conscience and intentionality. A rabbi remarked in the discussions that if AI is conscious then it will be enslaved.
The Theological View on Creation and Intent
The mystical, cosmic Creator brings to Creation the understanding that we lack, knowing its own intentions and those of its creatures, including what moral, emotional, psychological, philosophical, or physical underpinnings inform the thoughts behind their acts.