Rep. Don Beyer (D-VA) has his eye on artificial intelligence, along with a bipartisan group of lawmakers who recognize the profound impact of the technology.
He has long recognized AI’s potential — noting its development goes back decades — but understands the average American is only now coming to grips with its implications.
“It begins with the idea of machine learning, which is the idea of putting lots and lots of data into our computers,” Beyer said during WTOP’s Road to Responsible AI event. “And looking for connections, for correlations, or even causes that we may not be able to see as human beings because there’s just too much data out there.”
Beyer noted one of his favorite “factoids” is that more information was generated in 2023 than in the first 2,000 years since the birth of Jesus Christ all together.
But gathering all that information into computers comes with a lot of responsibility.
Beyer pointed out that a lot of ethical issues come up in the area of healthcare, for example. People don’t want a machine deciding whether they qualify for a type of medicine or surgery, for instance.
“We still want a doctor or human being to make help us make that decision together,” Beyer said. “And so, a lot of what we’re working on in Congress right now is trying to make sure that human beings are the final agents in the use of AI.”
The AI Caucus in Congress
Congress is often criticized for its need to catch up to developing trends and technologies. But there is an AI Caucus that includes Beyer and members of both parties.
Beyer noted that before the November presidential election, lawmakers met and endorsed more than a dozen bills related to AI.
“The expectation is every one of us would co-sponsor every one of those bills,” he said. “So we’re trying to get ahead of it, and I hope and expect that we will pass up to a dozen meaningful AI bills this year — put them on the president’s desk.”
Beyer contrasted that with how Congress failed for many years to address critical issues involving social media. He noted that for decades, lawmakers did little, except make it more difficult to sue social media companies.
Also, it wasn’t until relatively recently that lawmakers pressed technology companies to be more transparent with their algorithms and how they point users, including children, to specific content.
Beyer said lawmakers are trying to make sure AI is more transparent.
“We’re trying to make sure that the database models as they move forward, do you understand what what’s in the data, where did the data come from?” Beyer said. “And that you really test it to make sure that the results coming out are actually true.”
AI developers have continued to deal with what they call hallucinations, which can create incorrect or nonsensical outputs.
Coming AI legislation
Beyer is a sponsor of the AI Foundation Model Transparency Act, which is now being considered by Congress.
He said it’s directed at AI products and their makers, like Anthropic, Google ,Microsoft and OpenAI.
“The first thing I want to know is, what are you training your data on?” Beyer said. “Where is this data coming from, so that we have some sense of security that it wasn’t 6 trillion words off the internet.”
Secondly, the legislation seeks information on how companies test their data.
Also, how does the product reach its conclusion? “Right now, there’s so much of AI that seems to be magic,” Beyer said.
Lawmakers want the AI models to not only provide suggestions but to provide information about why the models are making those suggestions.
While lawmakers want transparency and accountability, Beyer said the legislation can’t be overbearing. He points to what the European Union has done, which he said is “widely seen as very prescriptive, very regulatory.”
The goal of the House bill he said is to create regulations that have “a light touch. We want people still to be creative and feel they have a lot of freedom to innovate.”
Looking ahead into AI’s future
Beyer said he’s encouraged by the progress that has been made on AI bills that have moved forward even though none of them on their own is “life changing.”
“We weren’t naive enough to think we were going to build some huge superstructure of legislation right away,” he said. “Let’s do it little by little and try to get it right.”
One of the areas lawmakers are looking at is “deep fake” ads that have the potential to play havoc with political campaigns.
Among the efforts is to require a disclaimer if a political ad uses AI. But Beyer said it can get complicated and the Federal Election Commission has been unwilling so far “to pick up that gauntlet.”
“But Congress is getting close to trying to do something meaningful,” he said.
As for his constituents, when he talks to them about AI, they have both curiosity and some fear.
He said the biggest fear for many is the possibility that it will cause people to lose jobs by being replaced by computers.
Beyer said he understands that concern but overall feels a great deal of optimism about how AI can positively transform the world.
“My great optimism is from what it is doing for health care,” he said. “Our kids, our grandkids are going to be living to 110 to120 years old. The breakthroughs on health care — every single day — are really exciting for us.”
Discover more articles and videos now on WTOP’s Road to Responsible AI event page.
