Members of Congress typically aren’t the tech savviest bunch, but dozens of lawmakers say artificial intelligence is now a part of their life.
“It’s good for research,” Republican Rep. Mark Alford of Missouri told NOTUS of X’s Grok. “And I say, ‘Please cite your sources.’ As a former journalist, it’s very important to me to have sources.”
“I use it as I use a calculator. I’m not a mathematician, and I’m not a researcher either,” Alford added. He said he recently used Grok to help him get up to speed for a congressional hearing with Secretary of State Marco Rubio. “I’m hip, for a 61-year-old guy,” he said.
Rep. George Whitesides, a California Democrat, said he uses Google’s Gemini for research as well.
“I’ve found it to be a great tool for summaries, if I just want to learn about something or if I have a slightly higher-order question — I use it all the time,” he said.
House Republicans included a provision in their reconciliation package that would take away states’ ability to regulate AI, while advocates have pressured lawmakers to put in place more guardrails around the technology. But Congress has been slow to respond, and the technology has been rapidly adopted: Since OpenAI released ChatGPT in late 2022, it has amassed over 400 million weekly users, and other products like Gemini have grown their user bases to over 350 million monthly users. Many lawmakers now count themselves among those users.
“I use ChatGPT, I use it for research, I use it for communications, I use it to better understand complex documents,” Rep. Eric Swalwell, another California Democrat, told NOTUS. “It’s great to take an input for something that’s very complex and ask, ‘Can you summarize this?’ But you have to validate that it understands and that the input is correct because, you know, it’s like garbage in garbage out.”
AI models commonly respond with false or “hallucinated” information to some queries.
Rep. Aaron Bean, a Florida Republican, says he just recently started using ChatGPT and that he has “an aim to use it more.”
“I’ve installed ChatGPT on my phone, so when I need a statistic for my speech or to verify something, I will type it in, but I need to do better,” Bean said. “I’m in Congress, and I need to tell the truth. So if I need to know what was the unemployment rate in 1978, it’ll tell me,” he said as he got his phone out of his pocket and searched the question on ChatGPT. (ChatGPT’s answer to NOTUS: “The average unemployment rate in 1978 was approximately 6.1%, according to data from the U.S. Bureau of Labor Statistics.”)
Democratic Sen. Ben Ray Luján says he’s found the free version of ChatGPT helpful not only for answering general questions but also for helping him with questions about his health.
“I had a stroke three years ago, so I’ll ask questions there and questions about food and things of that nature,” Luján said.
And Sen. Cory Booker told NOTUS that he commonly uses apps like ChatGPT and Perplexity as research tools.
“I use it as an alternative for what I used to use for Google search,” he said.
Staffers in Rep. Seth Moulton’s office use AI for internal tasks like research and information consolidation, spokesperson Sydney Simon told NOTUS. Simon said that staffers have clear guidelines on how to properly use AI assistance.
According to the most recent House statement of disbursements, 19 member offices have paid for a ChatGPT Pro subscription in the last year. ChatGPT is one of the only commercially available AI tools the House’s IT policy permits usage of, along with Claude Pro, Adobe Cloud AI tools and GPTZero, an AI text detector.
Reps. Nanette Barragán, Raja Krishnamoorthi, Randy Feenstra, Andrew Garbarino and Andrea Salinas are among those who have paid subscriptions. The House Administration Committee and House administrative offices like the sergeant at arms’ have also declared recent payments to OpenAI.
Even though members of both parties have happily adopted the technology to help them out with some of their tasks, practical and ethical concerns surrounding AI have caused some of them to consider how much they want to engage with it.
“I know what it’s capable of, and so I’m interested in seeing what we can do to regulate AI,” Sen. Mazie Hirono of Hawaii said. Hirono told NOTUS she doesn’t frequently use AI chatbots, but she is aware of what the technology can do.
“I know it has its uses, but at the same time, I think there are concerns about how our kids look at education,” Hirono added.
Democratic Rep. Robert Garcia says he has played around with AI, but it’s simply “not my thing.” He also said he has concerns about artists and their original work.
“I mean, people in my office might use it for some things like writing letters and things, but I personally don’t use AI,” he told NOTUS. “For me, where I think AI is the most concerning is in art. I don’t get why people are trying to take away the jobs of artists.”
Labor groups have raised concerns about how quick advancements in AI models might impact people’s jobs. Unions are also advocating for those in the film industry whose work may be threatened by increasingly realistic AI video creation models.
Republican Sen. Josh Hawley of Missouri, who has talked about his concerns regarding intellectual property and AI training, told NOTUS he doesn’t use AI nor does his staff.
“I see all these tech types say stuff like ‘delete all IP law.” Well, of course, that’s because they want to go take everybody’s stuff,” he said. “They want to take anybody’s copywriting material and not pay for it.”
Multiple news outlets and creative groups have taken up legal action over copyright infringement against some of the main AI developers due to the way they obtain content to train their AI models.
Democratic Rep. Jamie Raskin told NOTUS that AI could be a useful tool against misinformation, citing a dispute the congressman had with Elon Musk earlier this year, when Grok fact-checked a false claim shared by the tech billionaire.
“Up until they start manipulating it, I’ve got a lot of faith in artificial intelligence. I think you can use it to debunk a lot of propaganda disinformation online,” Raskin said.
But AI can also be a tool for creating misinformation campaigns, making it cheaper and easier to create deceiving text, images and videos.
Rep. Jasmine Crockett, a Texas Democrat, says she has mostly stayed away from AI due to misinformation on the internet.
“If anything, we’re constantly fighting AI because there is a ton of AI misinformation that is constantly written about me,” Crockett said.
Of course, not every lawmaker has gotten around to using AI chatbots.
“If you are talking about common AI tools, the only thing that I can think of that comes even close to that is that my iPhone suggests words and correct spellings,” Republican Sen. Susan Collins told NOTUS. “I don’t know whether that’s considered artificial intelligence or not.”
Others have not found the software particularly useful but have not been able to avoid it completely, either.
“I don’t have any term papers or anything,” Democratic Sen. John Fetterman told NOTUS when asked if he uses any AI platforms. “I wish I could AI some hair.”
“My experience with AI is on my iPhone and that really seemed underwhelming. And when I say, ‘FaceTime “G,”’ which is my wife, it then calls Jason Smith,” he added. ”So I don’t think Skynet is going to be our overlord. Siri keeps phoning my chief of staff when I’m trying to call my wife.”
—
Samuel Larreal is a NOTUS reporter and an Allbritton Journalism Institute fellow.