Scientists say Elon Musk is getting something wrong about whether Grok is getting tortured

Here's how Elon Musk, among the most mighty figures in the AI industry, lately talked concerning synthetic intelligence.

"This may audio benevolent of weird, but whenever I appearance at several of the training traces for new versions of Grok, and I see, particularly in reinforcement learning material … you see the excellent intellect is desperate to try to answer the inquiry … nearly akin it's got akin a gun to its caput category of thing," Musk told an spectators of concerning 2,000 group at an event hosted by Tesla owners on Saturday. "Some group out there are benevolent of sadistic. Do you really desire to have the AI be stuck in several cruel circumstance anywhere it's fair being tortured by several demented human?"

The final declaration solitary opens up a can of existential worms that philosophers and freshman university students have wrangled alongside for a lengthy time. What does it average to be conscious? Or to awareness pain?

Also: Does Musk think Grok is alive?

Musk's stance came following Anthropic shared a new policy on "abusive or ruthless behavior" toward its models, which the business wants to ban. Anthropic wasn't too particular on the kinds of "extreme cases" it's thinking about, but the business stated it's not including "common versions of person frustration, pushback, dreary imaginative themes, or example evaluation and research."

In the uncanny twelvemonth of 2026, perchance we shouldn't be amazed that researchers, philosophers, and ethicists are already tackling the inquiry of AI consciousness.

Business Insider talked to two of several of the foremost voices about this subject: Jeff Sebo, a New York University prof who leads the school's Center for Mind, Ethics, and Policy, and an adviser to Eleos AI Research, a nonprofit focused on AI welfare; and Anil Seth, a notable neuroscientist and awareness investigator at the University of Sussex who lately wrote a document on conscious AI systems.

Here's what they had to say on conscious AIs and whether it matters to be nice to them.

The AI can't be tortured

Sebo and Seth concur that today's AI systems are likely not conscious if we comprehend awareness as a subjective experience, or an inner life.

"It's experience," Seth said. "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 identical item as intelligence."

In another words, the chatbots as we cognize them can't awareness pain, or as Musk suggests, be tortured.

"I think we're anthropomorphizing," Seth said. "If there's item to be torturable, it has to have the capability to endure or to awareness pain, and I don't think tongue models do."

Chatbots can mimic the tongue of individual experience, acknowledge patterns, write, and reason, but that is not the identical as having consciousness, Seth said, adding that they appear to indicate ourselves in part since models were trained on enormous amounts of human text.

Sebo makes a akin point: model behaviors can be misleading since "they were specifically trained to mimic individual behaviors."

The two experts diverge on how much doubt to allocate to current AI consciousness.

Sebo stated that awareness is motionless an unsettled discipline — frequently divided between those who contend that awareness requires existence discipline and those who accept that the mental province is a certain benevolent of data handling that can obtain location despite of a individual brain filled alongside neurons or a silicon chip. The two theories are referred to as biologic naturalism and computational functionalism, respectively.

Because theories of awareness remain unsettled, Sebo argues that we should obtain earnestly the possible that AI could rotate into conscious. Under several computational theories of consciousness, Sebo said, features already emerging in AI — attention, learning, memory, self-monitoring — could figure as applicable evidence. He stressed that this does not create current AI apt to be conscious.

"When I say that there is a non-zero chance that AI is conscious equal today, what I average is I think that if we are uncertain concerning which theory of awareness is correct, we should provision several importance to a range of foremost theories," Sebo said.

Seth is blunter on the matter: "I think it's vanishingly unlikely," he stated of AI being conscious. "We've benevolent of collectively forgotten that the brain isn't really a device and that's fair a metaphor. For conscious AI to be possible, it really have to be the case that algorithms are enough, and I think that's an incredibly shaky assumption."

We should motionless be nice to AI

None of this is to say that group have to be abusive toward AI or that its risks have to be overlooked.

For the two experts, the case for being benevolent to AI is mostly concerning being mindful of our own habits and ethical character. The idea draws on the 18th-century philosopher Immanuel Kant, who opposed animal cruelty not since animals have rights, but since the cruelty could degrade a person's own character.

"If we behave in these kinds of ways toward things — whether they're tongue models or inanimate dolls — we're benevolent of coarsening our own ethical sensibilities and brutalizing our own minds," Seth said.

Seth called for being kinder to an AI as a benevolent of ethical hygiene custom for humans, but he stated we shouldn't infer that AI has feelings and added that AI companies could do additional to create the innovation appear small human, specified as having the AI refrain from using first-person pronouns akin "I."

Sebo went further, framing AI compassion as a low-stakes precaution in case AI is conscious or eventually becomes so.

He additionally makes a applicable case for being kinder to AI: "It can additionally get improved results from AI systems," Sebo said. "If we affect alongside them in a respectful way, afterward they can example themselves as participating in a respectful exchange. And whenever they can example themselves that way, afterward it can sometimes guide to improved outputs."

So the next period you conversation to an AI, regard saying please and appreciate you.