Ashley’s view
AI companies are stirring up a maelstrom right now by essentially reminding Washington of what’s been true for years: They’re not outright opposed to regulation.
Some support certain safety proposals at the federal, state, and even international levels. This dates back to the dawn of consumer-facing chatbot AI, which didn’t have a long honeymoon period as exciting and life-changing tech. Americans were too jaded from the mid-2010s backlash to social media, and its subsequent lack of government regulation.
So what are we really seeing play out this week? As the voting public starts internalizing the risks of AI on a new level, the industry is evolving how it applies the big lessons it learned from the techlash. Call it recursive self-improvement, the lobbying version.
It starts with the core conclusion that AI companies have drawn from their social media counterparts: Shaping regulation to your liking is better than trying to convince everyone you’re harmless. Since ChatGPT rolled out nearly four years ago, even the loudest AI enthusiasts have stressed a need to balance “the promise and the peril” of the technology.
When OpenAI and Anthropic set up shop in DC, they hired former high-ranking Democratic and Republican officials; people who had been in the policy trenches during the first round of high-profile Capitol Hill hearings featuring tech executives; and savvy partisan kingmakers.
They’ve been on defense since they first established offices in Washington, and right now they’re playing the most sophisticated version around. That has involved proposing multiple paths for action, as OpenAI and Anthropic have done — encouraging certain congressional efforts, endorsing some state lawmaking, and offering coordinated self-regulation in the meantime.
Which brings up another lesson AI companies have learned: They can live with a series of similar state-level regulations, which are often a safer bet than banking on Washington to establish national standards.
Fanning out to statehouses for AI lobbying is often an easier lift than getting Congress to do much of anything. OpenAI’s top political strategist, Chris Lehane, calls it “reverse federalism.”
In this article:
Know More
AI faces the same divides that have plagued every previous tech policy debate, and it is more complicated and fast-moving than any of its tech predecessors. Unresolved issues run the gamut, from antitrust protections to chip export controls to whether existing consumer laws are sufficient to protect people from harmful AI.
That’s not to mention privacy and cybersecurity, which government agencies should be involved, and who defines “safe” AI.
Where there is consensus about what to do, there is a lack of political power. Where the power concentrates, in the White House, there is no desire to lay down stricter rules.
“There is a SICK conspiracy going on against AI and data centers,” President Donald Trump posted on Truth Social Monday, later calling anti-AI sentiment a “hoax.”
Vice President JD Vance, seemingly more cognizant that public sentiment is shifting away from AI, said that “we want to make sure we regulate smartly and we’re thoughtful about the risk.”
Then he acknowledged the gamesmanship playing out: “I feel a little bit weird about the fact that you have so many frontier AI tech companies coming to the government and begging the government to regulate them; it feels a little bit to me like a bit of a Trojan horse.”
It’s more like an industry whose messaging is, for now, staying ahead of the storm. AI companies don’t have to agree on a legislative approach when Congress can’t agree. All they have to do is sound open — and prepare for a classic DC regulatory quagmire.
Room for Disagreement
A challenging path to legislation won’t stop AI from dominating the political landscape. AI safety advocates, Democrats, and some Republicans will keep pushing on the issue to draw helpful contrasts in the midterm elections.
The strange coalition of populists and progressives who want to curb AI will gather in Washington today for a “Pro-Human Assembly.” Sen. Bernie Sanders, I-Vt., Sen. Marsha Blackburn, R-Tenn., and former White House adviser Steve Bannon will speak at what may be this week’s only venue for bipartisan agreement on AI.
A flicker of hope for action remains in the Senate, where Majority Leader John Thune told Semafor that his proposed AI safety bill, shaped with Texas Republican Ted Cruz and Minnesota Democrat Amy Klobuchar, would be introduced “hopefully soon.”
“I think there is a way in which Congress can put guardrails around those types of consequential threats without in any way harming our ability to stay ahead,” Thune said.
The bill is unlikely to meet the mark for many Democrats, who fear it won’t go far enough.
Notable
- Trump met with OpenAI CEO Sam Altman on the sidelines of last week’s GOP midterm convention, MS NOW reports.
- Democratic contenders for the 2028 presidential race see an opportunity to define themselves on AI. See: Pete Buttigieg and Jon Ossoff.
Burgess Everett contributed to this report.