Sen. Angus King was in the middle of a hearing this week with U.S. Secretary of Energy Chris Wright when he pulled out his phone to consult with Claude, an artificial intelligence chatbot that has surged in popularity in the United States recently.

King wanted details about wind and solar energy capacity before he pressed Wright on the Trump administration’s decision to cancel renewable energy projects around the country, and Claude, built by the company Anthropic, instantly delivered.

“You have to be careful with it, particularly when it’s talking about analysis. But for data, it’s very useful,” the independent senator, who caucuses with Democrats, told NOTUS. “I use it all the time.”

King isn’t the only AI-curious member on Capitol Hill, though at age 82, he’s certainly one of the oldest. Many Democrats are warming to AI in a personal and professional capacity despite deep concerns in their party about its impact on job security, the environment, human relationships and society writ large. Progressive critics of the industry, for example, have called for major regulations cracking down on AI and gone on the attack against construction of massive energy-hogging data centers used to power it.

In interviews on Capitol Hill this week, over a dozen Democratic senators described how they are actively experimenting with AI chatbots, most commonly Claude, in their daily lives and for help with official duties. Some rely on them in a casual way, using AI as a souped-up search engine to do research, draft memos and speeches, organize their schedule, and even plan their family vacations.

Others have relied on AI for more complicated tasks. Sen. Adam Schiff used it to draft a living trust for him and his wife. Sen. Brian Schatz used it to analyze the many nonprofit grant funding requests he reviews as part of his job on the Senate Appropriations Committee. Sen. Mark Kelly went so far as to use Claude to try to build his own stand-alone applications, which is perhaps not that surprising for a former astronaut.

Senate Democrats have stepped up their social media game in recent months, an acknowledgment they were badly outpaced by the omnipresent online persona of President Donald Trump and the AI-fueled social media put out by the White House. Sen. Cory Booker, who leads online engagement efforts for his caucus, has coached his colleagues on better ways to use social media. The 46-year-old senator is now also pushing them to adopt AI.

“To not use AI as a part of a larger communication strategy would be malpractice, and I am going to continue to push for my colleagues to constructively engage these tools in order to better convey their policies,” Booker told NOTUS.

Sen. Andy Kim, meanwhile, said he envisions AI playing a role in building a more responsive government, such as providing better constituent services on Capitol Hill.

“Everybody needs to get deeper in on this,” Kim, 43, told NOTUS. “Are there things that I’m concerned about? Of course, but I think that that shouldn’t make us so adverse to not try some things out.”

Even Sen. Chris Murphy, who has advocated for strict regulations on artificial intelligence to protect children from loneliness, sexual abuse and self-harm, acknowledged leaning on Claude to help him find the right analogy or choice of words.

“I’ll sometimes use Claude as a thesaurus, or, you know, help me craft a phrase that I’m just not landing,” Murphy, 52, said. “I see the benefit as much as I’m definitely afraid.”

In the House, AI use has been more widespread.

Rep. Raja Krishnamoorthi has used an AI deepfake of himself to raise awareness of the technology, while former Rep. Jennifer Wexton used an AI-generated voice clone of herself to give a speech on the House floor in 2024 after she lost her voice due to a neurological disease. Rep. Jake Auchincloss, an early adopter of the technology, also read a speech written by AI on the House floor in 2023 to show the capabilities of AI models.

Auchincloss told NOTUS that AI has taken on a bigger role since then in his work and in his personal life. His office built a “JDA bot” (per the congressman’s initials), that has been fed every public writing that Auchincloss has ever produced to keep track of the congressman’s positions, though he doesn’t use it to write his speeches.

He also said that AI reads his kids bedtime stories.

“I have three little kids, and I’ll do prompt engineering for a good story. I’m like, ‘I want it to be for a 6-year-old, and I want it to be about helicopters,’” Auchincloss, who is 38, said. “They love it, they’ll tell me, ‘Dad, tell us a story from your phone!’”

Like with most issues in Congress dealing with technology, younger lawmakers are often quicker on the uptake and more knowledgeable than their older colleagues. That’s particularly evident in the Senate, where the median age (64.7) is on the cusp of retirement for the vast majority of workplaces in America. Congress is also known for being slow to adapt to innovations, relying on outdated systems and badly lagging behind the private sector.

“A few years ago, back when I don’t know if it was Alexa or something, I did like, like, maybe three sentences of dialogue back when it was first starting, and that freaked me out,” Sen. Jeff Merkley, 69, recalled, referring to the voice-activated assistant launched by Amazon in 2014.

“I’m gonna be careful because I don’t trust them completely,” added 81-year-old Sen. Dick Durbin when asked if he planned to use the technology.

There are exceptions to the age divide on AI.

Independent Vermont Sen. Bernie Sanders, at age 84, has been one of the loudest critics of unregulated AI adoption, warning that AI and robotics could cripple the workforce while benefitting corporations and the wealthy, not to mention his fear regarding AI girlfriends. Last month, Sanders released a video featuring his peculiar “conversation” with Claude, where he spoke to a suspended phone running the chatbot at a desk in his office. He posed questions to Claude about data privacy, data center development and AI’s impact on politics

“You can’t understand what you’re talking about unless you get a sense of what the technology is about,” Sanders told NOTUS.

Of course, lawmakers often rely on staff to brief them on key issues around new technologies. But there isn’t a unified policy in the Senate about the use of AI by staffers themselves, and Senate offices are free to set their own guidelines. Some of them prohibit it outright, while others are actively in the process of developing AI handbooks for staffers, acknowledging that AI isn’t perfect and can sometimes give users incorrect information.

“I’m saying like, like, use your best discretion, but you as the employee, are personally responsible for whatever information you put in front of the senator, right?” said Sen. Elissa Slotkin, 49, of her office policy on AI.

Schatz, 53, also said he was careful about how he used it.

“The problem is, it could still be wrong, right?” he said. “Sure, also, people can be wrong, but it’s wrong more often.”

While Democrats are getting more comfortable with AI on Capitol Hill, they’ve been more cautious about adopting it in their campaigns and in their campaign messaging. The Democratic National Committee, for example, has barred staffers from using ChatGPT and Claude, according to Axios.

Meanwhile, some Republican candidates have used AI to generate likenesses of their Democratic opponents in negative ads. The National Republican Senatorial Committee also released an AI video last month of James Talarico, the Democratic nominee for Senate in Texas, reading his past tweets on race and transgender rights. The video contained a small disclaimer in the right-hand corner, saying it was “AI-generated.”

Republicans relying on AI is “one of the reasons why I downloaded Claude myself and started messing around with it, so I could better understand the capability,” Kelly, 62, told NOTUS, lamenting how Democrats fell behind Republicans in past campaign cycles on adopting social media and participating in podcasts.

Sen. Kirsten Gillibrand, who chairs the Democratic Senatorial Campaign Committee, said Democratic candidates ought to feel free to use AI however they wish.

“You can do some pretty good, funny ads, but that’s up to the candidates. They decide how they want to portray their campaign,” the 59-year-old senator said.

Booker told NOTUS: “I think that we are in a moment of technology acceleration like we’ve never seen before. I do not want my party to be caught flat-footed and not engaging, not experimenting, and innovating in this space.”