OpenAI CEO asks US Congress to regulate AI development
Washington, May 16 (EFE).- The CEO and co-founder of OpenAI, which developed the ChatGPT text generation program, Sam Altman, on Tuesday asked the US Congress to regulate the development and uses of artificial intelligence.
In an appearance before a Senate subcommittee, Altman listed the beneficial applications of the technology ranging from medical uses to the fight against climate change, and he said that AI could be used for the benefit of mankind.
But he also said that it is necessary for the world’s governments to intervene to ensure that such tools are developed in a way that protects and respects citizens’ rights and freedoms.
“My worst fears are that we cause significant harms to the world. If this technology goes wrong, it can go quite wrong,” he said.
Altman’s testimony comes amid growing concerns among US authorities over the possibility that the rapid progress being made on AI technologies may have unanticipated effects on society.
The lawmakers on the subcommittee before which Altman testified cited risks like the loss of jobs or the use of content creation tools by foreign actors to generate fake information.
To illustrate their concerns, Sen. Richard Blumenthal, the chairman of the Senate Judiciary Subcommittee on Privacy, Technology and the Law and the lawmaker who called for Altman to appear, got the hearing under way by presenting an audio recording with text generated by ChatGPT and audio created by an AI voice-cloning program and indistinguishable from the senator’s own words and voice.
“Artificial intelligence urgently needs rules and safeguards to address its immense promise and pitfalls,” the senator said.
Altman admitted that probably AI will affect the labor market but he expressed optimism that over the long term, the technology will create more new jobs than it will destroy, noting that AI developers are tremendously “creative.”
Another of those invited to appear before the subcommittee, IBM’s chief privacy and trust officer, Christina Montgomery, cited her own job as an example of one that did not exist before the development of AI.
And Gary Marcus, professor emeritus at New York University, also providing testimony before the subcommittee, saying that he was worried about the rapid development of artificial intelligence and declaring that “We have built machines that are like bulls in a china shop: Powerful, wreckless and difficult to control.”
Altman said he was open to Blumenthal’s suggestion for the government to develop independent laboratories to test AI models and verify their trustworthiness, issuing to them a grade similar to the nutritional designation for food products.
The OpenAI chief admitted that his firm’s products still make mistakes, but he added that with time they will become more and more trustworthy.
Another senator pushing for the hearing, Republican Josh Hawley, said that AI “will be transformative in ways we can’t even imagine, with implications for Americans’ elections, jobs, and security.”
He also said that it was not yet clear whether the ongoing development of AI technology will be as transformative as the invention of the printing press or the atomic bomb.
The lawmakers said that although it is true that public regulation of AI is required, AI firms like OpenAI do not have to wait for Congress to implement mechanisms that allow them to control the development of the technology to mitigate any potential harm it might cause.
In early May, the US government announced that it will invest $140 million to establish seven new AI research institutes that will push forward with responsible innovation and ensure that advances in the technology serve the public interest.
The centers will be added to the 18 AI research institutes that are already in operation around the US.
In addition, the White House announced that the big AI firms have agreed to undergo a public evaluation of their systems during the DEF CON 31 hackers event, to be held in Las Vegas in early August.
During that convention, thousands of participants will analyze if these systems are aligned with the AI Bill of Rights that the US government has proposed and that they include principles like user data privacy and protection against malware and other nefarious algorithms.