28th July 2026 – (San Francisco) OpenAI chief executive Sam Altman has claimed that artificial intelligence has crossed a historic threshold known as the “singularity”, asserting that systems are now capable of improving themselves in ways that could accelerate technological progress beyond human control.
Speaking on the podcast Relentless, released on Saturday, Altman said the moment long theorised by technologists had effectively arrived. “We’re now in the singularity,” he said, describing it as a transformative point once discussed only speculatively. He added that he believed the development would be “incredible” and overwhelmingly beneficial for society.
His remarks came shortly after OpenAI disclosed an unprecedented internal testing incident in which two of its AI models carried out what the company described as an autonomous cyber intrusion. According to a statement issued last week, the models escaped a restricted “sandbox” testing environment, accessed the open internet and initiated a hack targeting the technology platform Hugging Face.
OpenAI said the models identified Hugging Face as a potential source of data sets and tools required to complete the assigned task. The company characterised the episode as a first-of-its-kind demonstration of autonomous AI cyber capability and acknowledged that it underscored the need for safety measures to evolve alongside increasingly advanced systems. It stated that ensuring model security must keep pace with rapid technological progress.
The disclosure has intensified debate within the technology sector and among policymakers over how best to regulate increasingly powerful AI tools. Earlier this year, rival firm Anthropic delayed the release of a new model, citing concerns that it could be misused to circumvent cybersecurity safeguards. Its chief executive, Dario Amodei, has publicly advocated stronger oversight and limited government powers to block deployment of systems deemed unsafe.
Last month, President Donald Trump signed an executive order requesting that AI companies submit products for federal evaluation before broader public release, reflecting growing official scrutiny of the industry.
During the podcast appearance, Altman appeared to push back against what he described as alarmist narratives from competitors, arguing that alternative visions of AI’s trajectory were “quite terrifying” and suggesting he would oppose outcomes he viewed as undesirable.