Skip to content

Content Library

Back to Articles

Article June 10, 2026

Who Holds the AI Brake as Banks Face Growing Risks?

Who Holds the AI Brake as Banks Face Growing Risks?

Anthropic co-founder Jack Clark's warning that the AI industry lacks a "brake pedal" has reignited debate over who should control the development of increasingly powerful artificial intelligence. For banks, the question is no longer theoretical. The chall
  • Anthropic co-founder Jack Clark warned the AI industry has a gas pedal but no brake pedal
  • Anthropic has proposed a coordinated mechanism allowing AI development to pause if risks become unmanageable
  • Debate continues over whether governments, regulators, or AI developers should control AI limits
  • PRA Chief Executive Sam Woods described AI-related cyber threats as the most significant risk facing banks
  • Regulators expect increasing disruption as AI systems become better at identifying software vulnerabilities
  • Banks must strengthen governance, model validation, cyber resilience, and human oversight
  • AI is rapidly becoming critical infrastructure across financial services
  • Success will depend on balancing innovation with accountability and control

The artificial intelligence industry has become accustomed to moving at extraordinary speed.

Yet when Anthropic co-founder Jack Clark warned that the sector has a "gas pedal" but no effective "brake pedal", his comments resonated far beyond Silicon Valley and into the boardrooms of the world's largest financial institutions.

Clark's concern centers on the possibility that AI systems could eventually begin developing increasingly powerful successors with limited human intervention.

Anthropic recently called on leading AI developers to consider a coordinated and verifiable mechanism to pause development if technological progress begins to outpace society's ability to manage the associated risks.

The company warned that increasingly autonomous systems could eventually create conditions in which humans struggle to maintain effective oversight.

Speaking to the BBC, Clark summarized the problem bluntly. "You want the option to be able to take your foot off the gas and put your foot on the brake. Right now, it's like the AI industry has a gas pedal, but it doesn't have a brake pedal."

The warning has reignited a fundamental question that governments, regulators, and corporate leaders have yet to answer satisfactorily: who should be responsible for limiting the scope and impact of AI?

Anthropic argues that responsibility cannot rest solely with governments or individual companies. Instead, the company has proposed a coordinated framework involving major AI developers, policymakers, and regulators.

Critics, however, contend that private technology firms should not determine the pace of technological progress, particularly when they have significant commercial interests at stake. Others argue that only governments possess the legitimacy and authority to establish enforceable boundaries.

For banks, however, the debate over who controls the brake is secondary to a more immediate challenge.

Artificial intelligence is already embedded in customer service, fraud detection, software development, cybersecurity, compliance monitoring, credit decisioning, and risk management. The technology is no longer experimental. It is becoming operational infrastructure.

That reality has prompted increasingly direct warnings from regulators. Sam Woods, Chief Executive of the Prudential Regulation Authority, recently described AI-related cyber threats as the most significant risk facing banks.

He warned that frontier AI models are becoming increasingly capable of identifying vulnerabilities in technology systems, creating the potential for substantial disruption across the financial sector.

Woods said it is "reasonable to expect quite significant disruption" from advanced AI systems and called on firms to improve their cyber defenses, accelerate software patching, and strengthen operational resilience.

His comments highlight a growing concern among supervisors that AI is simultaneously creating opportunities and expanding the attack surface available to criminals and hostile actors.

As AI models become better at discovering software flaws, generating convincing phishing campaigns, and automating cyber attacks, banks face an escalating arms race between defensive and offensive capabilities.

The challenge is compounded by the rapid adoption of AI within financial institutions themselves.

Many banks are already deploying generative AI tools to improve productivity and reduce costs. Yet governance frameworks have often struggled to keep pace with the technology's evolution.

Industry experts increasingly argue that responsibility for controlling AI risk must be shared. Governments must establish clear regulatory expectations.

Events you may be interested in

View all of our upcoming events

Sign Up to Our Newsletter

Sign up

Connect Plus

Unlock more with Connect Plus membership

This content is available to Connect Plus members.