OpenAI’s CEO, Sam Altman, has issued a grave warning about an upcoming “world-shaking” cyberattack in 2026.
In a recent discussion with Axios, Sam Altman elaborated on how highly sophisticated AI models possess the capability to transform the realms of cybersecurity and biosecurity.
The co-founder of Axios directly asked whether a catastrophic cyber incident such as a cyberattack could realistically occur within 2026. Altman responded, “I think that’s totally possible. Yes.”
Altman noted that AI models are advancing significantly.
At this stage of development, they present considerable risks to cybersecurity.
When discussing next-generation AI models, he indicated that they will be more dangerous than ever, and once such capabilities become available in open source, the dynamics will change.
“I suspect that within the next year, we will encounter substantial threats that we must address from cyber, as these models are capable and will continue to improve their capabilities in the near future,” Altman added.
The CEO of the company behind ChatGPT also highlighted the escalating risks linked to biosecurity.
He mentioned that the models are becoming exceptionally proficient in advanced biology, leading to remarkable advancements, such as the potential to cure diseases that have plagued humanity for centuries.
Anthropic warns of cyberattacks by future AI models
Earlier, experts warned that the upcoming surge of AI-driven cyberattacks would outpace those carried out by human hackers in the past.
This was the warning conveyed by AI firm Anthropic in a recently leaked blog post, stating that its upcoming AI model, Mythos, along with similar models, can exploit vulnerabilities at an unmatched speed.
Moreover, it is not alone in this concern.
OpenAI also indicated in December that its forthcoming models present a “high” cybersecurity threat.
Read more: OpenAI warns upcoming models may pose high cybersecurity risk
