McNerney’s peers across the Atlantic also felt a lack of enthusiasm. Brussels had stepped up its efforts to regulate the technology in 2020, but there seemed to be little political momentum when Romanian lawmaker Dragos Tudlache of the European Parliament, who co-leads the AI effort, contacted the US caucuses. was.
That has changed. The overnight success of AI-powered ChatGPT has sparked a frenzy among lawmakers in Washington to draft new legislation to address the promise and dangers of the burgeoning sector. Tudrash witnessed the AI hype when he visited Washington last month and attended a bipartisan conference with OpenAI CEO Sam Altman.
“It’s a different atmosphere,” Tudrash said in an interview.
But working with rapidly evolving technologies requires a sophisticated understanding of the complex systems that underpin AI, which can baffle even experts. Compared to Silicon Valley’s exorbitant salaries, congressional pay caps make it harder to keep engineers employed, putting lawmakers at a disadvantage in tackling this goal. This goal has become more urgent as the European Union overtakes Washington. Just this week, strong AI laws are being pushed forward.
to catch up Members of Congress and their staff are calling for crash courses on AI. As Senate Majority Leader Charles E. Schumer (New York) prepares to unveil plans on how Congress can regulate AI on Wednesday, lawmakers are suddenly at a briefing with industry leaders. It has taken various measures, such as rushing in and summoning influential scholars to discuss it. Concern about emerging fields.
Gaps in legislators’ technical expertise create gaps in corporate profits. Executives with a desire to develop AI unhindered flock to Washington, hoping to help educate lawmakers and influence policy. Schumer said his office has met with nearly 100 outside experts, including “CEOs of AI companies, scientists, AI scholars, industry leaders of all stripes, and AI critics.” said Microsoft president Brad Smith and Tesla were among those included. CEO Elon Musk.
The allure offensive has led some consumer advocates to fear that lawmakers will force the industry to make its own rules, which some executives openly recommend. Former Google CEO Eric Schmidt argued in an interview this spring that industry, not government, should set “reasonable boundaries” for the future of AI.
“There’s no way a non-industry person can understand what’s possible. It’s too new, it’s too difficult, it doesn’t have the expertise,” Schmidt told NBC. “No one in government can get it right. But industry can roughly get it right.”
Other industry leaders are taking different tactics, blitzing Congress with their vision of how the U.S. government should regulate their companies.Altman in May He held private meetings and dinners with lawmakers, where he demonstrated how ChatGPT could create speeches for the floor to entertain them. Smith gave lawmakers a lesson about the tech stack behind generative AI models like ChatGPT, including the computing infrastructure and applications. And Smith recently unveiled a blueprint for AI regulation in a speech in Washington attended by six lawmakers.
Smith said the stereotype that Congress doesn’t understand technology (reinforced by high-profile gaffes at major tech hearings) is “outdated” and calls for Congress to keep pace with AI progress. He added that he was “optimistic” about the ability of
A more formal education is provided by regular information sessions. Senate and House leaders hosted discussions on AI with MIT professors, reviewing the basics of how AI works and examining technology challenges such as how it exacerbates existing prejudices.
During Tuesday’s briefing with Massachusetts Institute of Technology (MIT) professor Antonio Tolarba, hosted by Schumer’s office, some lawmakers asked basic questions about how AI learns and where to get data, the session said. said former Senator Jackie Rosen (D-Nevada), a computer programmer who left the conference. quick.
Speaking at an April briefing arranged by Speaker of the House Kevin McCarthy, Republican, Calif., MIT professor Alexander Madley said, “They’ve put a lot of time and effort into understanding AI. I am working hard,” he said. Madry has since taken a leave of absence and is working at OpenAI.
Sen. Ted Cruz (R-Tex.) expressed skepticism about these efforts, suggesting that his colleagues’ technical acumen was desperately lacking.
“Honestly, Congress doesn’t know what the hell they’re doing in this space,” Cruz said, wearing earbuds, while attending the Politico Tech Summit via videoconference. “This is a facility [where] I think the median age in the Senate is about 142. The Senate is not a tech-savvy group. “
Senator Mark R. Warner (D, Va.), who previously worked as a venture capitalist, brought in researchers and industry leaders to speak with senators after Schumer’s congressional briefing. . According to Warner spokeswoman Rachel Cohen, his guests include Microsoft Chief Science Officer Eric Horvitz, Center for Security and Emerging Technologies Executive Director Dewey Murdick, and Deputy National Security Advisor Anne.・Various experts, including Mr. Neuberger, were reportedly involved.
“Many of us all go through different learning curves,” Warner told reporters on Tuesday.
The increase in AI briefings and attendance is a big change for Congress. In Congress, a handful of lawmakers (some of whom have degrees in computer science) have long struggled to attract the attention of their peers. Cruz said Congress hosted the first public hearing on AI in 2016 and chaired the conference. House members launched the AI Caucus in 2017, and senators launched a similar effort in 2019.
The rise of generative AI has finally awakened interest in such efforts. Rep. Mark Takano (D-California) said AI would affect “every jurisdiction in Congress,” and lawmakers drained funds in partisan battles at the Capitol’s tech think tank, Tech Assessments. He argued that it was necessary to respond by reviving the station. in the 1990s. Takano, along with Commerce Commission member Senator Ben Ray Luhan (DN.M.), plans to introduce a bill to fund the office next month.
“What Congress lacks is a pool of expertise that can be more forward-looking, quicker and more responsive,” Takano said. “We want to have expertise that is not tainted or tied to commercial interests.”
Some argue that concerns over the Capitol’s lack of technical expertise are exaggerated, with lawmakers already introducing legislation that could address most of the issues of generative AI, including data protection and algorithmic audit bills. claims to be.
“Congress’s job isn’t necessarily to get the nook and cranny of every technology it regulates,” said Anna Lenhart, who worked on technology policy for Rep. Lori Trahan (D-Massachusetts). says. “Their job is to understand the impact, risks and benefits of technology on society.”
Legislators can ask the Government Accountability Office and the Congressional Research Service for technical evaluations. Zach Graves, executive director of the American Innovation Foundation, said GAO’s resources have increased in recent years, resulting in better preparation for technology hearings like Altman’s.
“They obviously did more homework,” Graves said.
Still, some fear that the recent spate of corporate lobbying on AI has brought lawmakers uncomfortably close to regulated industries.
Sarah West, managing director of the AI Now Institute and former senior adviser, said unlike the clashes with CEOs of Facebook and Google, the friendly hearings between lawmakers with Mr. said it reflected how effective such intimate events were. On AI at the Federal Trade Commission.
West said executives like Google’s Schmidt are fueling the perception that AI is too difficult for Congress to understand.
It’s a “convenient narrative of taking accountability out of the hands of people to whom the public has entrusted it and into the hands of a profiting industry,” she said.
