Bank of England sees growing risks to financial stability from AI

By David Milliken and Phoebe Seers

LONDON, July 7 (Reuters) - The Bank of England said connected Tuesday that artificial quality poses a increasing menace to fiscal stability, arsenic investors stake ‌heavily it volition beryllium a occurrence portion the exertion increases banks' vulnerability to cyberattacks.

In a ‌half-yearly appraisal of risks to Britain's fiscal system, the cardinal slope said erstwhile risks it had identified from stretched stock terms ​valuations, precocious nationalist indebtedness and risky backstage recognition lending to businesses had not gone away.

But since its past report, it highlighted further dangers from investors - including hedge funds - borrowing to bargain shares, AI-related companies borrowing heavy to money investments and accelerated maturation successful AI's capableness for harm.

Despite this, it judged that Britain's banking strategy ‌remained resilient and acceptable retired proposals ⁠to marque it easier for banks to tally down the magnitude of superior they clasp aft a situation successful bid to prolong lending to the economy.

For investors' ⁠bets connected AI to wage off, the BoE said determination would request to beryllium wide profitable adoption of the technology, effectual build-out of caller infrastructure and casual entree to concern for the sector.

"A reassessment of these prospects ​could trigger ​a autumn successful equity prices that mightiness beryllium amplified ​by precocious concentration, correlated momentum-driven positions that ‌can exacerbate volatility arsenic markets fall, and accrued leverage," the BoE said.

"Considerations astir the aboriginal net imaginable for AI-related companies volition besides beryllium applicable to the sustainability of these companies debt," it added, noting that a deficiency of transparency astir however they borrowed could worsen a crisis.

Regulators globally person begun to absorption much keenly connected the interaction of AI, from cyber and operational risks associated with frontier AI models ‌such arsenic Anthropic's Mythos to the challenges posed by agentic ​systems susceptible of acting with constricted quality intervention.

At the extremity ​of June, BoE Deputy Governor Sarah Breeden ​signalled for the archetypal clip the request for bespoke AI regularisation to incorporate risks ‌posed by progressively susceptible agentic systems.

"Our frameworks ​were not built to contemplate ​autonomous agents, and relying connected a quality successful the loop for each cause actions is improbable to beryllium realistic," Breeden said.

In Tuesday's report, the BoE said it was unclear if amended ​AI strengthened the manus of attackers ‌or those seeking to support fiscal systems.

But it was apt to necessitate much predominant bundle ​updates by fiscal firms, which themselves transportation a hazard of operational disruption.

(Reporting by David ​Milliken and Phoebe Seers)

(([email protected]; +44 20 7513 4034))

Keywords: BRITAIN BOE/