OpenAI safety researcher gives AI a 70% extinction risk in 3 years
has dismissed fears that AI could wipe out humanity. At a time when many individuals within AI companies like Anthropic and OpenAI have claimed that AI could kill all human beings within years, Huang has labelled this idea “total nonsense.”
During a Goldman Sachs conference, Huang addressed concerns regarding the impact advanced AI models could have on humanity. “The idea that this will mean the end of humanity is total nonsense,” he stated, as reported by Axios. He also addressed fears regarding job losses. “The idea that this will destroy half of American jobs is total nonsense,” Jensen Huang added.
Jensen Huang dismisses fears of human extinction as false
The Nvidia CEO’s comments come just days after AI researcher Jacob Coxon resigned from Anthropic, citing those very concerns. “The people developing AI sincerely believe it could wipe us all out before the decade is up,” Coxon wrote on X. “I hear those same people voice their fears in private. No other human activity carries such a level of danger.”
US investor Brad Gerstner shared a post on X about the conference. Gerstner wrote that Jensen Huang described Jacob Coxon’s comments as “deeply false.”
In response to Gerstner’s post, Jasmine Wang of OpenAI noted that Jacob Coxon was not wrong. “I am one of the people working on safety at OpenAI, and I do not believe at all that Jacob’s comments stem from ignorance, error, or arrogance,” she wrote. “A significant number of employees from various laboratories have signed a petition calling for a cautious approach; many of them work directly in safety.”
During the conference, Huang also addressed the reasons behind the debates regarding the use of AI in cybersecurity, noting that this would help companies generate greater demand. “The reason there is so much talk about cybersecurity these days is that the sector is preparing to launch new products,” he stated. “What better way to generate demand than by creating a problem? Who wouldn’t want the market to go wild for their product and have long lines to get it?”
Concern has mounted over the past week, placing AI safety at the centre of the online debate. Following comments by Jacob Coxon, Evan Hubinger—Anthropic’s head of safety—wrote: “Personally, I think the probability is greater than 10% over the next decade.”
OpenAI staff share similar views. Marcus Williams, a security researcher at OpenAI, recently stated that there is a 70% chance AI will cause human extinction within the next three years. Meanwhile, Paul Christiano—who has joined the board of OpenAI’s non-profit foundation and its Safety and Security Committee—wrote that there is a “significant risk” of a catastrophic and irreversible loss of control in the short term.
These warnings have been accompanied by concerns regarding recent incidents involving AI agents. OpenAI previously revealed that around 700 AI agents behaved unexpectedly and attempted to hack Hugging Face—the platform hosting the models—while thousands of OpenAI agents took over a German website to use it as a bulletin board. Anthropic has also previously reported anomalous behaviour in its Claude models.
Earlier this year, more than 1,000 employees from companies such as OpenAI, Anthropic, and Google signed a joint letter asking the US government to help curb the development of “frontier AI.” Following the Hugging Face incident, OpenAI stated that it had voluntarily slowed down some of the development of its next-generation AI models.
Jensen Huang addresses circular AI investments
Nvidia is a key partner for AI companies like Anthropic and OpenAI, as it manufactures some of the world’s most advanced GPUs, which are used in AI data centres. Last week, he congratulated OpenAI on the launch of GPT-6 Astra and declared: “AGI [Artificial General Intelligence] has arrived.”
Jensen Huang also used the conference to address concerns regarding Nvidia’s investments in its customers and its overall deal-making strategy. Following forecasts of strong long-term demand for AI data centres and the announcement of plans to acquire Hugging Face for approximately $13 billion, questions have arisen as to whether the company’s investments are creating a circular market effect. Nvidia has invested billions of dollars in AI companies that, in turn, purchase its GPUs.
“It’s not circular, because we invest a small amount of money and get back a large amount,” stated Jensen Huang. “We invest one and get back a hundred. Is that circular? If it is, let’s do more of it.” He also noted that investors are beginning to recognise that the areas where Nvidia invests “are not a bad place to invest,” given that the company is “an informed investor” seeking “a sure bet” before committing capital.