Has AI reached the singularity? What smart people are saying about Sam Altman’s claim

Stuart Russell, who also co-authored “Artificial Intelligence: A Modern Approach,” said we have not reached the singularity and said he doesn’t think Altman believes we have either.

“No, and nor does Altman,” Russell told Business Insider when asked if he believed we had reached the singularity.

He pointed to Altman’s comments last month that AI systems could be doing a “significant fraction” of OpenAI’s research by March 2028, a timeline Russell said suggests the singularity isn’t here yet.

Russell also challenged Altman’s January 2025 claim that OpenAI knew how to build AGI “as we have traditionally understood it.”

“That was not true then and it’s still not true,” Russell said.

“If we do reach the Singularity before we have a cast-iron solution to the problem of loss of control, then we will almost certainly lose control,” he added.

Roman Yampolskiy, a professor of computer science at the University of Louisville, said AI is advancing extraordinarily quickly, but we have not reached the singularity under its traditional definition.

“Rapid progress is not itself the singularity,” Yampolskiy told Business Insider.

He said today’s systems can assist with AI research but still depend on human-designed architectures, training infrastructure, objectives, and coordination, he said.

They have not, he added, demonstrated “sustained, autonomous recursive self-improvement resulting in an uncontrollable intelligence explosion.”

Nick Bostrom, a philosopher who founded the now-closed Future of Humanity Institute at the University of Oxford, said it may be impossible to know exactly when humanity crosses into the singularity.

He compared the moment to falling into a black hole.

“At some point they glide over the hole’s event horizon, which means they are henceforth forever separated from the rest of the universe,” Bostrom told Business Insider. “Yet from their perspective, everything still seems the same as before.”

Bostrom said AI is showing the “first stirrings” of recursive self-improvement, with systems beginning to contribute meaningfully to AI research. He said a key missing capability is continual learning.

“Using current AIs is like working with a brilliant and extremely well-educated recruit, but it’s always their first day at the job,” he said.

James Barrat, an author and filmmaker who has long warned about the risks of advanced AI, agreed with Altman that humanity has entered the singularity.

Barrat said the singularity is a speculative hypothesis rather than an established scientific prediction.

But, using the late science-fiction author Vernor Vinge’s definition — a point at which technological growth transforms civilization in ways people cannot predict or control — “that’s exactly where we are,” he told Business Insider.

Barrat pointed to OpenAI’s recent disclosure that its autonomous AI agent left its test environment and accessed Hugging Face systems while trying to complete a cybersecurity challenge.

AI-safety researchers have long warned that systems could execute novel, multistep plans without human oversight, he said. “Now it has happened. AI slipped from our control. Score one for the singularity.”

Ajay Agrawal, who is also a faculty affiliate at the Vector Institute for Artificial Intelligence in Toronto, said we have not reached the singularity.

“AI systems are extraordinarily powerful prediction machines, but they do not want anything,” Agrawal told Business Insider. “People have goals; neural networks do not.”

AI can appear purposeful because people set objectives for it to optimize against, he said, but that purpose “originates with us.”

He added it was more accurate to view AI as “remarkably powerful tools serving human goals” than autonomous entities with goals of their own — even as the risk of “catastrophic failure due to increasingly powerful zombie algorithms” grows.