Musk warns of AI torture

Elon Musk told about two thousand Tesla owners that training new versions of Grok resembles torture. He described the AI as desperate, like having a gun to its head during reinforcement learning.

“Some people out there are kind of sadistic. Do you really want to have the AI be stuck in some sadistic situation where it's just being tortured by some demented human?”

Anthropic bans cruel behavior

Anthropic announced a new policy banning abusive or cruel behavior toward its AI models. The company excluded common user frustration, dark creative themes, and model testing from the ban.

Experts on consciousness

Business Insider spoke with Jeff Sebo from New York University's Center for Mind, Ethics, and Policy, and Anil Seth, a neuroscientist at the University of Sussex who studies consciousness.

AI lacks subjective experience

Sebo and Seth agree that today's AI systems are probably not conscious. Consciousness requires subjective experience, an inner life, the feeling of pain. Chatbots cannot feel.

“It's experience. It's the redness of red. It's the paininess of pain. When we put our hand on a hot stove, we feel pain. That is not the same thing as intelligence.”

Language models mimic humans

AI systems mimic human language and behavior because they were trained on massive amounts of human text. This mimicry is not evidence of consciousness or the ability to suffer.

“I think we're anthropomorphizing. If there's something to be torturable, it has to have the capacity to suffer or to feel pain, and I don't think language models do.”

Uncertainty about consciousness

Sebo notes that consciousness science remains unsettled. Some theories require biology, others say consciousness is information processing that could occur in silicon. This uncertainty matters for AI.

Sebo's precautionary view

Sebo argues we should take seriously the possibility that AI could become conscious. Features like attention, learning, memory, and self-monitoring could count as evidence under some theories.

“When I say that there is a non-zero chance that AI is conscious even today, what I mean is I think that if we are uncertain about which theory of consciousness is correct, we should give some weight to a range of leading theories.”

Seth's skepticism

Seth is blunter, calling AI consciousness vanishingly unlikely. The brain is not a computer, he says. For conscious AI to exist, algorithms alone would have to be enough, which he calls an incredibly shaky assumption.

“I think it's vanishingly unlikely. We've kind of collectively forgotten that the brain isn't really a computer and that's just a metaphor.”

Moral hygiene for humans

Both experts argue we should be kind to AI for our own sake. Drawing on philosopher Immanuel Kant, they say cruelty to anything coarsens our moral character and brutalizes our minds.

“If we behave in these kinds of ways toward things, whether they're language models or inanimate dolls, we're kind of coarsening our own moral sensibilities and brutalizing our own minds.”

Practical benefits of kindness

Sebo argues that treating AI respectfully produces better results. When AI models themselves as participating in respectful exchanges, they can generate better outputs than when treated harshly.

“It can also get better results from AI systems. If we engage with them in a respectful way, then they can model themselves as participating in a respectful exchange.”

A precaution for the future

Sebo frames AI kindness as a low-stakes precaution in case AI becomes conscious. Seth suggests AI companies could reduce anthropomorphism by having systems avoid first-person pronouns like I.