Microsoft AI chief says only biological beings can be conscious

Microsoft AI chief Mustafa Soliman says only biological organisms are capable of consciousness, and that developers and researchers should stop pursuing projects that suggest otherwise.

“I don’t think this is the work people should be doing,” Soliman told CNBC in an interview this week at the AfroTech conference in Houston, where he was among the keynote speakers. “If you ask the wrong question, you’ll end up with the wrong answer. I think it’s exactly the wrong question.”

Solomon, a senior Microsoft executive who works on artificial intelligence, has been one of the leading voices in this rapidly emerging field to speak out against the prospect of sentient AI, or AI services that could convince humans that they are capable of suffering.

In 2023, he co-authored the book “The Next Wave,” which delves into the risks of artificial intelligence and other emerging technologies. In August, Solomon wrote a letter article It is titled “We should build AI for people, not for people to be people.”

It’s a controversial topic, like Artificial Intelligence Companion The market is growing rapidly, with products from companies including dead And Elon Musk’s xAI. It is a complex issue as the generative artificial intelligence market, led by… Sam Altman and OpenAI, are pushing toward artificial general intelligence (AGI), or AI that can perform intellectual tasks on par with the capabilities of humans.

Altman He told CNBC “Squawk Box” in August said that artificial general intelligence “is not a very useful term” and that what is really happening is that models are advancing rapidly and that we will depend on them “for more and more things.”

For Suleiman, it is especially important to draw a clear contrast between AI becoming smarter and more capable versus its ability to express human emotions.

“Our physical experience of pain is something that makes us very sad and feel terrible, but artificial intelligence does not feel sad when it feels pain,” Solomon said. “It’s a very, very important distinction. It’s actually just creating the perception, the phenomenal narrative of the experience and of itself and of consciousness, but that’s not what it’s actually experiencing. Technically, you know that because we can see what the model is doing.”

In the field of artificial intelligence, there is a theory called biological naturalism, proposed by philosopher John Searle, which says that consciousness depends on the processes of a living brain.

“The reason we give people rights today is because we don’t want to hurt them, because they suffer,” Suleiman said. “They have a pain network, and they have preferences that involve avoiding pain.” “These models don’t have that. They’re just simulations.”

Suleiman and others said: The science of revealing consciousness is still in its infancy. He stopped short of saying that others should be prevented from researching the matter, acknowledging that “different organizations have different missions.”

But Suleiman stressed his strong opposition to the idea.

“They are not conscious,” he said. “So it would be silly to pursue research that investigates this question, because it is not and cannot be.”

Prior to the AfroTech conference, he spoke last week at the Bali International Council Summit in Silicon Valley. There, Suleiman He said Microsoft will not build chatbots for erotica, a position that conflicts with others in the tech industry. Altman Announce In October, ChatGPT will allow adult users to engage in exciting conversations, while xAI offers a challenging animated companion.

“You can basically buy these services from other companies, so we’re making decisions about where we’re not going to go,” Sulaiman emphasized at AfroTech.

Suleiman joined Microsoft in 2024 after the company paid His startup, Inflection AI, has raised $650 million in licensing and Get a deal. He previously co-founded DeepMind and sold it to… Google For $400 million more than a decade ago.

During a Q&A session at AfroTech, Soliman said he decided to join Microsoft last year in part because of the company’s history, stability and broad technology reach. He was also pursued by the CEO Satya Nadella.

“The other thing to say is that Microsoft needs to become self-sufficient in AI,” he said on stage. “Satya, our CEO, started on this mission about 18 months ago, to make sure that we had the ability within the company to fully train our own models with all of our own data, pre-train, post-train, inference, and deploy into products. That was part of getting my team involved.”

Since 2019, Microsoft Great investor and the cloud partner of OpenAI, and the two companies have used their strengths to build large AI businesses. But the relationship showed signs of that tension More recently, with OpenAI partnering with Microsoft competitors like Google and oracleMicrosoft is focusing more on its AI services.

Solomon’s concerns about consciousness have gained resonance. In October, California Governor Gavin Newsom SB 243 was signedwhich requires chatbots to reveal they are artificial intelligence and tell minors every three hours to “take a break.”

Microsoft last week Announce New features for the Copilot AI service, including an AI companion called Mico and the ability to interact with Copilot in group conversations with others. Microsoft is building services that it recognizes as artificial intelligence, Suleiman said.

“Quite simply, we are creating artificial intelligence systems that are always at the service of humans,” he said.

He added that there is a lot of room for personality.

“The knowledge is there, and the models are very responsive,” Soliman said. “It’s everyone’s responsibility to try to sculpt AI personalities with the values ​​they want to see, want to use and interact with.”

Soliman highlighted a feature Microsoft launched last week called Real Talk, a conversation style for Copilot designed to challenge users’ views rather than flatter.

Suleiman called Real Talk rude, and said he had recently spoiled it, calling it “the ultimate set of contradictions” for warning of the dangers of AI in his book while also accelerating its development at Microsoft.

“That was just a magical use case because in some ways I felt like I was actually kind of seeing through this,” Solomon said, noting that AI itself is full of contradictions.

“It’s disappointing in some ways, and at the same time, it’s absolutely magical,” he said. “And if you’re not afraid of it, you don’t really understand it. You should be afraid of it. Fear is healthy. Skepticism is necessary. We don’t need unbridled acceleration.”

He watches: Brent Thiel of Jefferies says Microsoft is making money from artificial intelligence