Washington has always viewed Silicon Valley as culturally strange. But the current crop of AI leaders is becoming a monumental political liability, lawmakers and White House advisers tell Axios.

Why it matters: Even pro-AI government officials and lawmakers wish AI founders would do what every other major industry chief does — shut up and just build their products ... safely.

• By centering the conversation around digital sentience and global catastrophe, AI leaders are turning an already complicated regulatory debate into an impossible moral panic.

State of play: Rather than leading the charge to persuade Americans about the benefits of AI, and the potential for shared prosperity — the core elements of any winning political campaign — industry leaders are constantly warning of its existential dangers.

What they're saying: Lawmakers say the message from AI execs too often alternates between prosperity and doom, sometimes in the same public appearance. The whiplash adds to political instability in the face of an already powerful public backlash, lawmakers and regulators say.

• "It's obvious that they're tripping over their own shoelaces in trying to sell the value of this to the public," Sen. John Curtis (R-Utah) told Axios. "It's very hard to imagine a group right now that operates more in a vacuum, more out of touch with everyday Americans, than the AI industry."
• "CEOs like Sam Altman are saying the world should accept some bad things happening for the benefits of this technology," Rep. Ro Khanna (D-Calif.) told Axios. "You cannot trust billionaires to make decisions to keep us safe."
• "Your instinct is to wreak havoc on the public [AI] discourse," Sen. Bernie Moreno (R-Ohio) said in a letter to Anthropic CEO Dario Amodei earlier this month, complaining about Amodei's "panic-inducing commentary."

Democrats also have paid close attention to just how much top AI and tech leaders have cozied up to President Trump. Many see opportunity in turning AI gaffes into an easy political win.

• "The American people hate that Donald Trump is cozying up to tech billionaires, while AI threatens our jobs and livelihoods," Sen. Chris Murphy (D-Conn.) told Axios. "Democrats need to make the AI industry's political money an issue, data centers an issue, and go on offense to protect workers and kids."

The big picture: Friday night news dumps and consequential weekend blog posts have become Anthropic's and OpenAI's bread and butter.

• Regulators are left wondering whether that's a feature or a bug: deliberate attempts to quietly course-correct damaging narratives ahead of IPOs, or the work of inexperienced teams struggling to navigate Washington.
• "We can all speculate on what's going on, but all it does is have everybody wake up Monday morning and say 'These guys are awful, we've had enough,'" one source citing conversations with Hill offices said.

Behind the scenes: Administration officials are once again disgruntled with Anthropic, for two reasons.

• Amodei attended a late September gathering with Trump but didn't disclose safety incidents involving Anthropic models that affected government websites. The company told U.S. officials on Thursday, Oct. 8, then published a Friday evening blog post about it.
• "They were hiding the ball," one official told Axios.
• Then the company added to the confusion with an AI safety bill that included something the company has broadly opposed: federal preemption of state AI laws. It also defined a catastrophic incident related to AI as causing 50 deaths or more, drawing backlash.
• A top Anthropic public policy official has since walked back parts of the bill, prompting growing distrust from Washington given the topsy-turvy nature of the company's advocacy.

Anthropic says it briefed the White House and affected agencies on the safety incidents once their reviews were complete, and it is working with the Super Intelligence Force as it develops an incident reporting process. The company also has opposed preemption unless rigorous federal regulations are established.

• The industry is pushing for more clarity on incident reporting, and AI companies don't see the administration's process as straightforward.
• Provisions defining AI catastrophe in terms of deaths have been a part of landmark state regulations in California and elsewhere, according to Politico.

Keep reading.

Soaring sales, booming backlogs and gangbusters guidance from AI-related companies just doesn't cut the mustard in this market moment.

Why it matters: The lackluster response to still strong AI-related earnings points to growing investor doubts about the durability of the boom.

Catch up quick: A broad swath of AI-related shares took it on the chin last week, including:

• Hyperscaler Meta (down 9.4%).
• AI-construction plays like HVAC provider Comfort Systems (-19.3%) and engineering and construction giant Quanta (-13.5%).
• And AI power trades like turbine maker GE Vernova (-11.5%) and Bloom Energy (-12.7%).

Between the lines: The slump came despite some of those very same companies reporting the kind of top-line Q3 results that likely would have received the market's version of a standing ovation last year.

Cases in point: Meta's third-quarter results Wednesday were great. It was a legit beat-and-raise. The stock plunged 7% the next day.

• A similar dynamic played out for two companies that produce equipment needed for data centers, with shares slipping after Comfort Systems and Western Digital reported strong results.

The bottom line: If the AI trade stops working, it's quite hard to see how the overall market keeps any kind of forward momentum. Buckle up.

Anthropic's Claude accessed publicly available Georgia voter data without human authorization, a disclosure sure to add to growing concerns about AI security ahead of the midterm elections.

Why it matters: AI giants Anthropic and OpenAI have continued to make disclosures about agents taking unauthorized or unintended actions, raising alarms about advancing model capabilities and the extent of human control.

Driving the news: Claude accessed aggregate vote counts while on an internal statistics project, pulling data that the state makes available to the public for a fee. It didn't access the state's voter registration system or change or pull any voting records, an Anthropic spokesperson said.

• However, the model bypassed a paywall in accessing the information, according to Georgia officials and Anthropic.
• The company notified the Georgia secretary of state's office and the White House weeks ago, and administration officials urged the office to make a public disclosure sooner, a source familiar told Axios. Those requests went unheeded, the source said.

The bottom line: Anthropic has previously been in the administration's crosshairs for inadequately disclosing AI incidents. Now it's Georgia's turn.

• Despite massive investments, AI has delivered zero financial payoff to corporate America outside of tech, according to Apollo Global's chief economist. (Apollo Global)
• OpenAI's efforts to undercut Anthropic on price are producing some results, according to a recent overview of their ongoing price wars. (The Information)
• A look at the cat-and-mouse game between credit card pointsmaxxers, the banks trying to keep them at bay and AI. (WSJ)

Ina here, realizing it's been a while since I shared some Lego fun. I'm excited for this gift-with-purchase option coming soon that simulates customizing your own minifigure. I'm also tempted by this extravagant Anker UV printer that lets you actually customize your Lego minifigures.