Clueless Capitol Hill jolted as latest tech development deemed too dangerous for public

WASHINGTON — President Donald Trump and Sen. Ted Cruz (R-TX) are leading the charge to eradicate state artificial intelligence laws that Silicon Valley entrepreneurs say are stifling, but there’s a problem: There’s still no federal AI law to replace local tough-on-AI measures.

As Senate Commerce Committee chair, throughout this Congress, Cruz has preached the business-friendly gospel of preemption — a legal doctrine that makes state-passed laws subservient to federal statutes — around Washington.

“It certainly is one element that is an important priority of the president,” Cruz told Raw Story at the Capitol recently. “We'll see where the votes are in Congress.”

On the other side of the aisle, Democrats are baffled that Cruz is fighting robust local AI statutes passed by local elected officials in California, Utah, Colorado and 35 other states when there’s no federal measure to replace these state laws with.

“Oh, yeah, he would like Congress to not have any say in this,” Sen. Mark Kelly (D-AZ) told Raw Story. “In the absence of federal protections with things, you know, states certainly have a right to protect its citizens from whatever it happens to be.”

“Do you think we can afford, with AI, to go the laissez-faire route?” Raw Story asked.

“No, I think we need guardrails,” Kelly said. “The range of possibilities of how this technology is going to turn out is pretty broad. We've got to narrow it down to a place where it's a win for the American people and workers.”

As of now, the AI boom — and boon — has surely been a win for tech companies and their soaring stock prices, but lawmakers in both parties are now fighting to make sure average Americans become casualties of today’s AI revolution.

“Advances in AI continue to be exponential”

Most members of Congress are clueless when it comes to AI.

Few lawmakers had even heard of deepfakes — AI-generated content that mimics before distorting reality — before a deepfaked robocall of President Joe Biden was blasted to thousands of New Hampshire primary voters in 2024.

In the Senate, where the average age is about 64 years old, lawmakers were so behind on the technology sweeping the globe that then-Senate Majority Leader Chuck Schumer made the rare decision to close the chamber’s doors to the public for three all-senator AI briefings and nine AI Insight Forums during the 118th Congress.

Many of the AI advances senators were briefed on last Congress are now obsolete. And the tech isn’t slowing down.

While there hasn’t been much attention paid to AI on Capitol Hill since those closed-door Senate briefings, news of Anthropic's new Claude Mythos AI model has awoken the sleeping giant that is Congress in recent months.

The company refuses to release its Mythos model publicly, because executives say it’s too powerful and dangerous for the public.

That’s, in part, because the company says it’s expert at exploiting security vulnerabilities in the internet as we know it, i.e. blowing through firewalls and making mincemeat out of today’s cyber security measures.

That’s why Anthropic alerted the federal government of its surprising advancements and is only making the Mythos model available to a handful of cyber security companies enrolled in its new Project Glasswing.

“There's no doubt the advances in AI continue to be exponential,” Sen. Cruz said. “And that has both enormous productivity gains but also very real downside risks.”

While next-generation artificial intelligence is here, Congress is still debating what to even debate.

“Do you see a sweeping package or more targeted bills?” Raw Story asked the Commerce Committee chair.

“We're in the process of figuring that out and seeing which elements would have the maximum positive impact and at the same time command sufficient support to be able to pass,” Cruz said.

Top of Cruz and Trump’s list is preemption — where state AI laws are forced to play second fiddle to the feds — despite there being no federal AI law, just executive orders and enforcement actions at federal agencies.

The federal preemption effort has had a cooling effect on local legislatures.

In March 2024, Utah was one of the first states out of the gates with its Artificial Intelligence Policy Act — a forward-looking artificial intelligence measure aimed at extending consumer protections to the AI-era while also instituting some narrow disclosure requirements on tech companies.

But in 2025, under pressure from Silicon Valley firms, Utah lawmakers wound some of the law’s disclosure requirements back.

Same in California, New York and Colorado, which were all quick to heap new disclosure, governance and transparency requirements on AI, only to then wind some of their own new regulations back in the face of Silicon Valley backlash.

“I'm normally a states' rights girl”

With Washington gridlocked, state legislatures have, once again, been ground zero for new AI regulations — with AI proposals now introduced in all 50 state capitals and passed in 38 of them — but all their work may be for naught.

“I'm normally a states' rights girl,” Sen. Cynthia Lummis (R-WY) told Raw Story. “But on this, I get the point that there needs to be a certain base level of protection for artificial intelligence.”

Lummis, who was a member of the far-right Freedom Caucus when she served in the House, says artificial intelligence is an interstate commerce issue, “so the federal government has jurisdiction.”

“Correct me if I'm wrong,” Raw Story pushed. “But I don't really see a big AI bill coming or even a small one?”

“Well, therein lies the challenge,” Lummis said. “So even for those of us who agree that it's appropriate to have some base level of federal preemption, the question becomes what's the level? Where is the level? And that is not easy, and it's not really all that agreed upon.”

Not even within her own Republican Party, where critics of the nation’s technology class are up in arms over efforts to let Silicon Valley “regulate” itself.

Yesterday’s artificial intelligence anxiety is now today's AI reality.

“It's here,” Sen. Josh Hawley (R-MO) told Raw Story. “It's here.”

While the AI revolution continues upending industry — with an estimated 85,000 job cuts due to AI so far in 2026 alone — Congress keeps dithering.

“There needs to be legislation on this,” Hawley said.

For his part, Hawley’s teamed up with Sen. Richard Blumenthal (D-CT) to introduce the GUARD Act — which is aimed at shielding children from chatbots — and the Artificial Intelligence Risk Evaluation Act, which would “establish an Advanced Artificial Intelligence Evaluation Program within the Department of Energy for assessing AI risks and informing federal oversight.”

“It's pretty modest, our bill is, honestly,” Hawley admitted. “But we need something here to make sure that we're tracking the national security risks, because that's a big one.”

For now, it’s up to AI companies like Anthropic to police themselves and their ever-evolving new AI models.

While that’s good enough for Trump and Commerce Chair Cruz, Silicon Valley self-policing just won’t do with AI.

“I'm not sure every AI company would have been this responsible in terms of timing of release,” Sen. Mark Warner said. “For me, it raises a huge amount of questions about, you know, who can we rely on before this gets released into the wild?”

“And yet here on the other side of the aisle,” Raw Story pushed, “they're still talking about preemption of state laws with a federal bill.”

“Over a long term basis, I've always said, you know, well, preemption makes sense, but only if you've got strong guardrails,” Warner said.

“We don't even have a federal law yet,” Raw Story pushed.

“We don't have a federal law on privacy!” Warner bemoaned. “We don’t have a federal law on social media.”

Without even basic guardrails on technology firms, many are braced for disaster, especially when it comes to the Pentagon and Defense Secretary Pete Hegseth.

“I don't trust him with this technology,” Sen. Andy Kim (D-NJ) told Raw Story. “Well, then the question is, well then what? And that's what we need to answer.”

“What are we actually trying to accomplish?”

The contemporary Congress has mastered new media, at least when it comes to their reelection efforts, but that’s about it.

“I worry that, you know, we've just consistently seen Congress fall short of actually tackling cutting-edge technology,” Kim lamented.

Kim says history’s now repeating itself with AI.

“I've been here in the Senate now about a year and a half, you know, we've yet to really have that type of depth and understanding of what we should be doing when it comes to AI,” Kim said. “So that's really the first and foremost that we need to do.”

While the White House and Trump’s allies on the Hill are fighting efforts to place guardrails around AI, Kim says they’re missing the mark, because there is no mark.

“We have to have a sense of, what are we actually trying to accomplish?” Kim said. “Then we can sort of figure out what the right level of government is and what interaction it is.”