OpenAI chief executive Sam Altman has urged the public to trust artificial intelligence companies to develop the technology safely, while acknowledging growing concerns about the risks posed by increasingly powerful AI systems.
Speaking at a Salesforce conference in San Francisco on Tuesday, Altman said people were justified in being concerned about AI because its capabilities had advanced rapidly.
He said OpenAI and other AI companies understood the seriousness of the risks and were confident they could keep safety measures ahead of technological development. Altman added that companies should slow down or stop development if they could not maintain adequate safety controls.
His comments followed renewed debate after a former Anthropic researcher warned that unchecked AI development could pose an extreme threat to humanity. Several AI experts and executives supported the warning, while others argued that some of the fears surrounding the technology were exaggerated.
Anthropic chief executive Dario Amodei has called for the pace of frontier AI development to slow and urged governments to introduce safeguards. He said on Tuesday that Anthropic was in discussions with other AI companies about stronger industry wide safety standards and checks.
Political figures have meanwhile questioned whether AI companies should be allowed to regulate themselves. Independent US Senator Bernie Sanders said decisions about AI should not be left to a small group of wealthy technology executives, while former Trump adviser Steve Bannon also criticised the industry’s ability to self regulate.
Nvidia chief executive Jensen Huang defended the role of AI companies in deciding when new systems should be released. He said safety should remain a priority but described it as an engineering challenge, arguing that companies should pause a release if they are not confident they can control the technology.
Meta chief executive Mark Zuckerberg also said AI laboratories had strong incentives to build safer systems, arguing that companies that failed to focus on AI alignment could fall behind and face liability if their models caused harm.
OpenAI executive Chris Lehane said the company was working with Anthropic and Google DeepMind on a voluntary effort to develop standards for frontier AI, with or without government involvement.
The debate comes as governments, technology companies and researchers continue to disagree over how much responsibility for AI safety should remain with the industry and how much should be overseen by regulators.
Leave a comment