Senators have finished their first secret briefing on artificial intelligence (AI), raising concerns over the risks posed by artificial intelligence (AI) technology and lacking a clear front on legislative plans to regulate the burgeoning industry. rice field.
Tuesday’s briefing was requested after Senate Majority Leader Chuck Schumer (New York) and others warned that lawmakers needed expertise in a rapidly developing industry, and Director of National Intelligence Avril Haynes. and top intelligence and defense officials, including the deputy secretary of defense. Kathleen Hicks and White House Director of Science, Technology and Policy Arathi Prabhakar brief Senators on the risks and opportunities posed by AI.
Senator John Kennedy (R-Louisiana) said, “AI has amazing potential to make our lives better,” before pausing.
“We must be killed first,” he added.
Congress and the administration have been scrambling to better understand the risks and benefits of generative AI in recent months, especially since the rapid adoption of OpenAI’s ChatGPT tool following its release in late November.
AI-powered chatbots such as ChatGPT and Google’s Bard, as well as video, audio and image-based tools, are increasing concerns over the spread of so-called hallucinatory false information or false information shared by chatbots. increase. Further risks of how the technology will be weaponized.
“One of the interesting things about this space at the moment is that it doesn’t feel particularly partisan, so we have a moment to capitalize on it,” said the work that called for a series of AI briefings. Senator Martin Heinrich (DN.M.), one of four members of the section, said.
But he said lawmakers need to understand how the technology works, including the dangers that can arise from its limitations.
“It’s very important to understand how these models work. It just predicts what sounds like an appropriate response. It’s like a statistics game that’s being played, but it’s a lot different than realizing that guardrails aren’t built in a lot of the time,” he added.
“These models are not made to tell the truth. They are made to tell something that sounds like a proper English response.”
Senator Elizabeth Warren (D-Massachusetts) said the large-scale language models that AI systems are trained on were not designed with accuracy in mind.
“This creates a real threat that AI can be exploited to sound sensible while perpetuating wrong answers one after another,” she added.
But lawmakers were divided on how, and even if, AI should be regulated.
Senator Chris Coons, Democrat, Delaware, said: “We are more concerned than ever that there are significant challenges before us and that the Senate needs to pass legislation to address them. “I’m done with the meeting,” he said.
Koons called the call the latest in a “series of constructive conversations” about AI risks and opportunities, but there is still bipartisan agreement on a path toward legislative proposals to regulate the technology. has not been obtained.
Senator Marco Rubio (R-Florida), the Republican head of the Senate Intelligence Committee, said the road to regulation will be difficult.
“One thing I am sure of is that I do not know of any technological advance that could have been rewound in human history. How do we build guardrails and practices around them to increase the risk and reduce its harm,” Rubio told reporters.
He also warned that lawmakers will face limits if they try to regulate the private sector that promotes the technology around the world.
“In terms of how the government uses it, or what some companies in the United States are doing, it’s possible,” Rubio said. “But unlike some technologies in the past, AI is not the sort of thing that is neither knowledge-based nor engineering-based. There is a possibility that other countries will develop it, so I don’t know if it will be solved from a global perspective.”
Rubio added that he was not opposed to regulation.
“I don’t have a particularly good knowledge of AI, and I can’t even understand what we’re trying to regulate. I’m not ready to give you an opinion on that, because I think we’re still learning.”
Senator Mazzy Hirono (D-Hawaii) described AI as “uncharted waters,” but thought Congress’ role would best address the technology in the realm of politics.
“Especially in the field of politics, we provide some parameters so that people can recognize if something is AI-generated content so they can recognize that it may not be trustworthy. But there are many other applications and uses for AI that I don’t think we fully grasp,” she said.
Some lawmakers have warned that Congress should be wary of overregulation that could hurt competition. A leading tech company in the space makes similar points.
“On the one hand, you don’t want to ignore the risks, and you don’t want to ignore the obvious impact of technology. state) said.
Nick Clegg, global president of Meta, urged tech companies to lead with transparency as they move forward with the adoption of AI tools, in an op-ed published in the Financial Times on Tuesday. . At the same time, he said, “the most dystopian warnings about AI are really about a technological leap, or a few leaps, beyond the current state of the art.”
“There’s a world of difference between chatbot-style applications of today’s large language models and super-large frontier models that could theoretically deliver sci-fi style superintelligence. , we are discussing the dangers we might encounter at the summit,” Clegg wrote.
“But there is time for both technology and guardrails to develop,” he added.
Tuesday’s briefing is part of Schumer’s plan for how lawmakers will tackle regulating a burgeoning industry, and comes alongside a series of expert forums Schumer has announced will be held later this year. will be held.
Schumer unveiled the AI regulatory framework last month. He founded a bipartisan group of senators consisting of Heinrich, Todd Young (Republican, Indiana), and Mike Rounds (Republican) to lead the issue with him.
Schumer’s proposed framework follows other voluntary guidelines on AI announced by the administration through the White House’s AI Bill of Rights Blueprint and the National Institute of Standards and Technology’s (NIST) AI Risk Management Framework. But the lack of strict government regulation has forced the tech industry to develop mostly voluntary guidelines.
Sen. Mark Wayne Mullin (R-Oklahoma) said he hopes to see more hearings and working groups made up of people “really interested” in the subject.
“We are not [know a ton about AI]. It was 30,000 feet in altitude and was kind of an introduction. It was very interesting. Basically everyone said: “Listen, technology has been around for a long time, but this is a new thing that is advancing rapidly.” , let’s figure out what’s wrong, ”he said.
“Parliament tends to overreact sometimes. Let’s not overreact yet, because there is a need here. But we also need to be careful here and not be taken advantage of by our enemies.”
