OpenAI CEO Sam Altman warns of an AI ‘fraud crisis'

CNN

5 days ago

OpenAI CEO Sam Altman says the world may be on the precipice of a 'fraud crisis' because of how artificial intelligence could enable bad actors to impersonate other people.
'A thing that terrifies me is apparently there are still some financial institutions that will accept a voice print as authentication for you to move a lot of money or do something else — you say a challenge phrase, and they just do it,' Altman said. 'That is a crazy thing to still be doing… AI has fully defeated most of the ways that people authenticate currently, other than passwords.'
The comments were part of his wide-ranging interview about the economic and societal impacts of AI at the Federal Reserve on Tuesday. He also told the audience, which included, representatives of large US financial institutions, about the role he expects AI to play in the economy.
His appearance comes as the White House is expected to release its 'AI Action Plan' in the coming days, a policy document to outline its approach to regulating the technology and promoting America's dominance in the AI space.
OpenAI, which provided recommendations for the plan, has ramped up its presence on and around Capitol Hill in recent months.
On Tuesday, the company confirmed it will open its first Washington, DC, office early next year to house its approximately 30-person workforce in the city. Chan Park, OpenAI's head of global affairs for the US and Canada, will lead the new office alongside Joe Larson, who is leaving defense technology company Anduril to become OpenAI's vice president of government.
The company will use the space to host policymakers, preview new technology, and provide AI trainings, for example, to teachers and government officials. It will also house research into AI's economic impact and how to improve access to the technology.
Despite Altman's warnings about the technology's risks, OpenAI has urged the Trump administration to avoid regulation it says could hamper tech companies' ability to compete with foreign AI innovations. Earlier this month, the US Senate voted to strike a controversial provision from Trump's agenda bill that would have prevented states from enforcing AI-related laws for 10 years.
Altman isn't alone in worrying that AI will supercharge fraud.
The FBI warned about these AI voice and video 'cloning' scams last year. Multiple parents have reported that AI voice technology was used in attempts to trick them out of money by convincing them that their children were in trouble. And earlier this month, US officials warned that someone using AI to impersonate Secretary of State Marco Rubio's voice had contacted foreign ministers, a US governor and a member of Congress.
'I am very nervous that we have an impending, significant, impending fraud crisis,' Altman said.
'Right now, it's a voice call; soon it's going be a video or FaceTime that's indistinguishable from reality,' Altman said. He warned that while his company isn't building such impersonation tools, it's a challenge the world will soon need to confront as AI continues to evolve. Altman is backing a tool called The Orb, built by Tools for Humanity, that says it will offer 'proof of human' in a world where AI makes it harder to distinguish what, and who, is real online.
Altman also explained what keeps him up at night: the idea of bad actors making and misusing AI 'superintelligence' before the rest of the world has advanced enough to defend against such an attack — for example, a US adversary using AI to target the American power grid or create a bioweapon. That comment could speak to fears within the White House and elsewhere on Capitol Hill about China outpacing US tech companies on AI.
Altman also said he worries about the prospect of humans losing control of a superintelligent AI system, or giving the technology too much decision-making power. Various tech companies, including OpenAI, are chasing AI superintelligence — and Altman has said he thinks the 2030s could bring AI intelligence far beyond what humans are capable of — but it remains unclear how exactly they define that milestone and when, if ever, they'll reach it.
But Altman said he's not as worried as some of his peers in Silicon Valley about AI's potential impact on the workforce, after leaders such as Anthropic CEO Dario Amodei and Amazon CEO Andy Jassy have warned the technology will take jobs.
Instead, Altman believes that 'no one knows what happens next.'
'There's a lot of these really smart-sounding predictions,' he said, ''Oh, this is going to happen on this and the economy over here.' No one knows that. In my opinion, this is too complex of a system, this is too new and impactful of a technology, it's very hard to predict.'
Still, he does have some thoughts. He said that while 'entire classes of jobs will go away,' new types of work will emerge. And Altman repeated a prediction he's made previously that if the world could look forward 100 years, the workers of the future probably won't have what workers today consider 'real jobs.'
'You have everything you could possibly need. You have nothing to do,' Altman said of the future workforce. 'So, you're making up a job to play a silly status game and to fill your time and to feel useful to other people.'
The sentiment seems to be that Altman thinks we shouldn't worry about AI taking jobs because, in the future, we won't really need jobs anyway, although he didn't detail how the future AI tools would, for example, reliably argue a case in court or clean someone's teeth or construct a house.
In conjunction with Altman's speech, OpenAI released a report compiled by its chief economist, Ronnie Chatterji, outlining ChatGPT's productivity benefits for workers.
In the report, Chatterji — who joined OpenAI as its first chief economist after serving as coordinator of the CHIPS and Science Act in the Biden White House — compared AI to transformative technologies such as electricity and the transistor. He said ChatGPT now has 500 million users globally.
Among US users, 20% use ChatGPT as a 'personalized tutor' for 'learning and upskilling,' according to the report, although it didn't elaborate on what kinds of things people are learning through the service. Chatterji also noted that more than half of ChatGPT's American users are between the ages of 18 and 34, 'suggesting that there may be long-term economic benefits as they continue to use AI tools in the workplace going forward.'
Over the next year, Chatterji plans to work with economists Jason Furman and Michael Strain on a longer study of AI's impact on jobs and the US workforce. That work will take place in the new Washington, DC, office.