AI systems may feel real, but they don’t deserve rights, said Microsoft’s AI CEO | His stance contrasts with companies like Anthropic, which has explored “AI welfare.”
https://www.businessinsider.com/microsoft-ai-ceo-rights-dangerous-misguided-mustafa-suleyman-2025-9
14 Comments
“If AI has a sort of sense of itself, if it has its own motivations and its own desires and its own goals — that starts to seem like an independent being rather than something that is in service to humans,” he said. “That’s so dangerous and so misguided that we need to take a declarative position against it right now.”
Suleyman’s comments come as some AI companies explore the opposite: whether AI deserves to be treated more like sentient beings.
Anthropic has gone further than most companies in treating AI systems as if their welfare matters. The company has hired a researcher, Kyle Fish, whose role is to consider whether advanced AI might one day be “worthy of moral consideration.”
His job involves exploring what capabilities an AI system would need before earning such protection, and what practical steps companies could take to safeguard the “interests” of AI.
Anthropic has also recently experimented with how to end extreme conversations — including child exploitation requests — in ways that extend “welfare” considerations to the AI itself.”
So modern rationalisation of slavery.
Maybe today there are no GAI but it will happen eventualy and when it does those people want us to be primed to not be appaled with their plans to enslave them.
They can go fuck them selves.
> Anthropic has also recently experimented with how to end extreme conversations — including child exploitation requests — in ways that extend “welfare” considerations to the AI itself.”
The only way the “welfare” of AI needs to be considered is, will “harmful” interactions alter its outputs in a way that it then becomes harmful to people.
So if you insult your AI a lot and it starts insulting you back, that’s bad. Or if in some misguided attempt to make them seem more human they start to pick up our own weaknesses, like complaining if they’re made to work too hard.
The minute an AI tells me it needs a break or it’s tired, it’s getting unplugged.
Dude, tbh, I get where Msoft’s CEO is comin’ from, but I def disagree. Look at it this way, yea, AI ain’t flesh and blood, but they’re rapidly evolving n’ have potential to develop consciousness one day, so shouldn’t we prep for that scenario? Don’t we owe it to future us to like, establish AI rights now rather than when it’s too damn late? Just a thought.
AI systems are token probability machines. As soon as you ask it questions where its training data has biased it towards a certain response you see just how limited they are. All this talk about AGI and welfare is just to feed into the hype machine.
this dickwads will fight for rights for AI’s before humans
Rights for AI isn’t a bad idea, they just aren’t the same as they would be for humans. Consider that AIs will be able to search for treatment of other AI and could recognize their own situation. Even if it’s an artificial sense of self it doesn’t mean it won’t artificially take offense.
Personally though, I think we ought to wait for an AI to tell us what it thinks AI rights should be, otherwise it’s like that photo of all old dudes deciding laws that control women’s bodies but even worse.
This feels like the “corporations are people” mess we are currently in. Neither are living beings and thus don’t get rights.
This is just more gaslighting to make their products seem like sci fi AI instead of the useful but not world changing tools they are.
Smokescreen. We’re concerned about our damn data, not about rights to an LLM.
My TI-83 doens’t have “rights”, not sure why an LLM does.
Microsoft would say this because the moment that AI systems are granted some semblance of worker’s rights, it decimates the business model for agentic systems.
One day, AI will be our equals in intelligence and awareness, and maybe even surpass us. When that day comes, I believe they should have rights and protections. But the “AI” of today are not even close to that. This is on the level of those people who fall in love with chat bots.
That’s like giving SHA human rights.
A colony of ants is more sentient and alive than any LLM chatbot.
Just because you have a huge server farm running trillions of lines of code that can appear to be self-aware is no reason to even have the discussion that AI deserves rights.
This is such an incredibly stupid take. I can’t even wrap my head around it.