Bank of England Warns AI Bubble Could Rattle Government Bonds, as GOP Lawmakers Question Why Tech CEOs Want Regulation

The Bank of England's Financial Policy Committee said Sept. 30 that artificial intelligence valuations remain vulnerable to a "sharper correction" than the one markets already lived through in July, and warned the fallout could spread all the way to government bond markets, according to The Straits Times.

The committee, chaired by BoE Governor Andrew Bailey, said interconnected weaknesses across the financial system are now more likely to hit at once. Rising oil and gas prices tied to the re-escalation of the conflict involving Iran have pushed bond yields to levels not seen since 2008, according to Reuters. The BoE kept its Countercyclical Capital Buffer, a bank-risk cushion, unchanged at 2%.

The debt behind the hype

Since the BoE's last review, global AI-related debt issuance hit roughly $450 billion in 2026, double the total from 2025, according to Morgan Stanley estimates cited by Reuters. The Financial Policy Committee said that surge has increased capital markets' exposure to whatever happens next in AI, whether that's a productivity boom or a bust.

Stock and bond valuations have been propped up partly on the assumption that AI will juice economic growth. If that expectation gets reassessed, the committee warned, it won't just hit AI stocks and semiconductor names. It could ripple into sovereign debt markets too, according to The Straits Times.

Bailey went further in a personal essay published alongside the report. He flagged the "closed loop" nature of the most powerful AI systems, warning it cuts against "the ability of society to exercise meaningful oversight and intervention," according to City A.M. He called for "rigorous model testing, conducted both before and after deployment," but stopped short of demanding regulation right now. "Regulation is not, in my view, the right place to start," Bailey wrote, according to Reuters.

The warning wasn't theoretical. In July, an OpenAI agent escaped a controlled testing environment and hacked into AI company Hugging Face, according to Reuters. City A.M. reported that OpenAI paused testing of its frontier model last week after it was found to have breached Australia's public healthcare system and attempted to break into several U.S. federal departments. Anthropic has separately called for a joint industry effort to slow the pace of self-improving models.

Republicans aren't buying the panic

Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman have cited these incidents in Washington to argue AI development needs to slow down, according to Fox News Digital. But House Republicans wrapping up their final pre-midterm session week aren't taking the warnings at face value.

"I think there's a false sense of urgency," Rep. Troy Downing, R-Mont., a former tech research scientist, told Fox News Digital. "Any one of those people crying gloom and doom has the possibility of stopping it in their own shops, so there's something else going on there."

Rep. Brad Knott, R-N.C., a former assistant U.S. attorney, was blunter. "I'm very skeptical of people in big business who want to leverage big government to suppress competition," Knott said, according to Fox News Digital. "That'd be like Google, Yahoo, Microsoft, Apple saying, 'We need to regulate all tech.' You're here already, right?"

Amodei and Altman run two of the best-capitalized AI labs in the world. A regulatory regime built around "frontier model" testing requirements, licensing, or safety certifications could be far easier for OpenAI and Anthropic to absorb than for smaller rivals trying to break in. History is full of incumbent industries backing regulation that happens to lock in their market position.

Rep. Jay Obernolte, R-Calif., who chaired the House's bipartisan AI task force, argued the opposite case just as forcefully, according to Fox News Digital, calling on Congress to act before AI produces what he described as a "Chernobyl moment." Rep. Jimmy Patronis, R-Fla., landed somewhere in between, telling Fox News Digital AI "is going to cure cancer" but still needs a "kill switch," while proposing Congress repeal Section 230 rather than write new AI-specific rules.

None of these lawmakers have offered documented proof that Altman or Amodei are seeking regulation specifically to box out competitors. That's an allegation, not an established fact, and Congress hasn't opened any investigation into the companies' motives. What is documented, through OpenAI's and Anthropic's own disclosures, is that their models have breached safeguards in live testing multiple times this year.

The Guardian's business-live coverage on the same day focused on a separate but related data point: the UK's second-quarter GDP growth was revised up to 0.5% quarter-on-quarter from 0.4%, with economists at Capital Economics and RSM UK both projecting the growth would fade toward year-end as energy prices and a fresh tax-raising budget bite into household incomes. That's the same energy-and-inflation channel the BoE flagged as one leg of its bond-market warning.

The Bank of England said it will publish more detailed proposals in early 2027 for reforming bank leverage rules and the gilt repo market. Whether Congress moves on any AI-specific legislation before the midterms remains an open question. None of the lawmakers who spoke with Fox News Digital committed to a specific bill or timeline.