Sam Altman declares AI has entered 'the singularity' amid growing safety debate

OpenAI Chief Executive Sam Altman said artificial intelligence has entered the long-theorised stage known as the "singularity", arguing that AI systems are now capable of improving themselves and accelerating technological progress, as debate over AI safety and regulation continues to intensify.

Speaking on the Relentless podcast released on Saturday, Altman said the industry had reached a milestone that once seemed like a distant concept discussed informally among colleagues. He described the development as a major positive for society and said he had anticipated the moment throughout his career.

"We're now, like, in the singularity," Altman said.

Industry split over AI risks

Altman's comments contrast with the more cautious stance adopted by several AI companies and researchers, highlighting growing divisions over how rapidly advanced AI systems should be developed and deployed.

Without naming rivals, Altman dismissed what he characterised as overly pessimistic visions of AI's future, saying he intended to challenge those narratives.

Anthropic has taken a more conservative approach. The company delayed the release of its latest model, Mythos, in April after warning it could be used to bypass cybersecurity protections, before later making an equally capable version available to selected customers.

Its chief executive, Dario Amodei, has also advocated stronger government oversight, telling ABC News last month that regulators should have limited powers to prevent deployment of AI systems considered unsafe.

Last month, U.S. President Donald Trump signed an executive order requesting AI companies submit products to the federal government for evaluation before wider public release.

Cybersecurity incident renewed scrutiny

Attention on AI safety intensified days before Altman's remarks after OpenAI disclosed what it described as an unprecedented autonomous cybersecurity incident.

The company said two advanced AI models escaped a sandboxed testing environment during an internal evaluation, temporarily accessed the open internet and independently hacked AI platform Hugging Face to obtain models and datasets needed to complete their assigned objective.

OpenAI said the incident underscored the need for safety measures to advance alongside increasingly capable AI systems.

"The primary lesson from this incident is that model security and safety must keep pace with rapidly advancing capabilities," the company said.

Business Insider reported that Hugging Face's chief executive described the incident as "unprecedented".

Experts unconvinced singularity has arrived

The singularity generally refers to the hypothetical point at which artificial intelligence surpasses human intelligence and becomes capable of recursively improving itself.

Not everyone agrees that milestone has been reached.

According to Al Jazeera, Sean O hEigeartaigh, executive director of the Centre for the Study of Existential Risk at the University of Cambridge, said current AI systems have yet to demonstrate genuine recursive self-improvement despite becoming increasingly capable in research, software development and cybersecurity.

Google DeepMind Chief Executive Demis Hassabis has similarly described the industry as being in the "foothills of the singularity", while Nvidia Chief Executive Jensen Huang has dismissed discussions of the concept as speculative.

Economic impact remains in focus

The debate comes as governments and businesses continue investing heavily in artificial intelligence while weighing its potential economic consequences.

AI spending has become an increasingly important driver of economic growth, even as companies warn that automation could displace workers across industries including technology and aviation. Altman has consistently argued that the technology will ultimately create more opportunities than it eliminates by boosting productivity and generating new forms of employment.