• Anthropic co-founder Jack Clark warned the AI industry has a "gas pedal" but no effective "brake pedal"
• Anthropic has proposed a coordinated mechanism allowing AI development to pause if risks become unmanageable
• Banks have largely responded by accelerating governance, oversight, and model-risk management efforts
• Regulators increasingly view AI-enabled cyber threats as a major risk to financial stability
• Financial institutions are strengthening human oversight and validation controls around AI deployments
• Debate continues over whether AI developers or governments should determine when development slows
• The banking sector remains committed to AI adoption but is demanding stronger safeguards and accountability

The global banking industry has reacted cautiously but sympathetically to warnings from Anthropic co-founder Jack Clark that artificial intelligence is advancing too quickly and requires an effective "brake pedal" before it becomes impossible to control.

His comments have intensified an already active debate within financial services about how to balance innovation with the growing risks posed by increasingly autonomous AI systems.

Clark's remarks followed Anthropic's proposal for a coordinated mechanism that would allow leading AI developers to pause the development of advanced models if risks begin to outpace society's ability to manage them.

The company warned that future systems could eventually design more powerful successors without direct human involvement, creating the potential for what researchers describe as recursive self-improvement.

While banks have not called for an outright halt to AI development, many senior executives appear to share Clark's concerns about governance and control.

Financial institutions have spent the past two years aggressively deploying generative AI across customer service, fraud detection, software development, compliance monitoring, and risk management functions.

Yet many are simultaneously strengthening internal controls amid fears that the technology is evolving faster than traditional model-risk frameworks can accommodate.

The banking sector's reaction has been shaped in part by growing regulatory concern.

Earlier this month, Sam Woods, chief executive of the Prudential Regulation Authority, described AI-related cyber threats as the most significant risk facing banks.

He warned that increasingly powerful AI models could expose vulnerabilities in financial institutions' technology infrastructures and force firms to accelerate software patching and cyber-defense efforts.

Industry observers say Clark's warning is reinforcing a view already taking hold within banking boardrooms: that AI governance can no longer be treated as a future issue.

Large institutions have expanded AI oversight committees, introduced stricter validation requirements, and increased scrutiny of third-party AI providers.

Many firms are also investing heavily in explainability, monitoring, and human-in-the-loop controls to ensure critical decisions remain subject to human judgment.

Some banking executives privately acknowledge that Anthropic's concerns are particularly relevant because AI is increasingly being used to write software code.

Clark recently noted that around 80% of Anthropic's code is now generated by its own Claude system, with that figure potentially rising further.

For banks, which rely on complex legacy technology environments, such developments raise questions about operational resilience, accountability, and cyber security.

Not everyone agrees with Anthropic's approach. Critics argue that calls for a coordinated slowdown could benefit large AI developers by making it harder for smaller competitors to catch up.

Others maintain that governments, rather than technology companies, should decide when and how restrictions are imposed.

Those arguments have also found support within parts of the financial sector, where institutions remain wary of allowing private technology firms to shape the rules governing critical infrastructure.

Nevertheless, the broader direction of travel appears clear. Banks remain enthusiastic adopters of artificial intelligence, viewing it as essential for improving efficiency and competitiveness.

At the same time, Clark's warning has strengthened the conviction among risk professionals that innovation must be matched by rigorous oversight.

As one of the world's most influential AI developers argues for the ability to slow progress when necessary, the banking industry is responding not by hitting the brakes, but by demanding a stronger steering wheel.

The result is likely to be a renewed focus on governance, resilience, and accountability as financial institutions seek to harness AI's benefits without losing control of the technology that increasingly powers their operations.