Washington, D.C., is quickly solidifying its role as the epicenter of the AI conversation. At Axios AI+ DC, lawmakers, technologists, and operators weren’t debating whether AI would reshape the economy; they were grappling with how fast and how unevenly it would do so, and who would shape the rules.
The shift is already underway: AI is no longer confined to product roadmaps. It’s showing up in legislation, labor markets, infrastructure planning, and increasingly, public anxiety.
What emerged across the sessions I attended wasn’t a single narrative, but a set of tensions. And those tensions will define what comes next.
AI and Government: Regulation, Competition, and Responsibility
U.S. Senator Mark Warner (D – VA)
Michael Kratsios, Director, White House Office of Science and Technology Policy
U.S. Senator Josh Hawley (R – MO)
There’s more agreement in Washington on AI than you might expect, and far less action than the moment demands.
Senator Warner pointed to real bipartisan alignment on AI safety and economic disruption. But alignment hasn’t translated into legislation, and the window to get ahead of impact, particularly job displacement, is narrowing.
At the same time, no one wants to be the reason the U.S. loses ground globally. As Warner put it, “we can’t stuff this back into the bottle,” especially with China in the picture.
The White House’s Kratsios leaned into that tension, outlining a pro-innovation approach that still attempts to protect consumers, especially as AI infrastructure begins to drive real costs, from energy to public resources.
Meanwhile, Senator Hawley made it clear that patience is running thin in some corners of Congress, particularly around child safety and AI systems interacting with minors.
The takeaway: Washington understands the stakes. The question is whether it can move fast enough to matter.
AI in Healthcare: One of the Few Places AI Is Already Working
Sandeep Dadlani, UnitedHealth Group EVP and Optum Insight CEO
While many industries are still experimenting, healthcare is already operationalizing AI in meaningful ways.
Dadlani didn’t frame AI as a future fix; he framed it as a present necessity for a system under pressure. Burnout, administrative overload, and fragmented patient experiences aren’t new problems. What’s new is the ability to actually address them.
Ambient AI, turning clinical interactions into structured, usable insights, is one of the clearest examples. It’s not flashy, but it’s effective: less time on documentation, more time with patients. At Inkhouse, we’re seeing this in practice with clients like Nabla, whose ambient AI, dictation, and real-time intelligence allow providers to focus more on care and less on collating and recording notes.
At the system level, AI is already identifying care gaps and improving outcomes, with measurable impact on cost. And the bigger shift is coming: a “digital front door” where the healthcare experience starts long before you walk into a clinic, and follows you after you leave.
The takeaway: Healthcare is the blueprint for what happens when we move past AI hype and focus on execution and results
AI and Sustainability: The Constraint No One Can Ignore
Joe Dominguez, president and CEO, Constellation
AI may be built in the cloud, but it can’t run without very real infrastructure.
Dominguez made the abstract tangible: data centers powering AI are already consuming energy at levels comparable to major cities.
That reality is forcing a shift in the energy conversation. Nuclear, long politically fraught, is re-emerging less as an ideal and more as a necessity: one of the few energy sources that is both reliable and low-carbon at scale.
At the same time, public concern is building. Energy costs, environmental impact, and the visibility of large-scale infrastructure are turning AI into a kitchen-table issue.
The takeaway: AI’s growth will be limited not by imagination, but by infrastructure. Energy is now a gating factor.
AI and the Fear Gap: The Technology Is Moving Faster Than Trust
Dina Powell McCormick, vice chair, Meta
Sarah Heck, head of public policy, Anthropic
Craig Martell, chief digital and AI officer, U.S. Department of Defense
If innovation is accelerating, trust is not keeping pace.
Powell McCormick acknowledged what many leaders skirt around: transformation at this scale is unsettling. But she also pointed to AI’s upside as an equalizer, lowering barriers and expanding access in ways previous technologies didn’t.
Anthropic’s data reflects that tension. Optimism exists, especially among users, but so does a clear expectation that the government should play a role in shaping outcomes.
Martell offered perhaps the most grounded perspective: AI isn’t intelligence in the human sense, it’s “statistics at scale.” Which means it’s powerful, but also fallible.
Layer in deepfakes, job anxiety, and unclear accountability, and you get what showed up repeatedly across sessions: a widening gap between what AI is and what people believe it is.
The takeaway: The biggest risk right now isn’t just misuse of AI, it’s misunderstanding it.
AI’s Future Will Be Defined by People, Not Just Technology
Axios AI+ DC made one thing clear: The technology cycle we once thought this was is actually a major structural shift.
The next phase of AI will be shaped less by model breakthroughs and more by policy decisions, infrastructure constraints, and institutions’ ability to keep pace with change.
As Dina Powell McCormick put it, we are witnessing a “transformation of humanity” in real time.
The next few years will determine who leads in AI, but more importantly, it will also determine how AI serves people.
