Debating AI regulation: AFC Soapbox Project at the Autumn 2024 UK Conference

Summary

At the AFC Autumn 2024 UK Conference, the AFC Soapbox Project ignited passionate debate with the motion: “This house believes that the auto and equipment finance industry needs clarity on UK AI regulation in order to properly leverage its value.”

Background

The EU has provided a framework for considering AI risk which classifies activities as unacceptable, high, limited, and minimal risk. Each class has different regulation and requirements for developing or using AI systems. This helps businesses consider how best to deploy AI without falling foul of future law and regulation but it can be difficult to practically apply the classifications to proposed use cases; it may drive activity towards low reward, low risk use cases rather than focussing on high value use cases where risk may be reduced [is this true?]; and it creates an uneven playing field disadvantaging EU firms over non-regulated competitors from outside the region.

The US by contrast seems to be following a different path – providing innovators with the freedoms they need to innovate; and waiting to deal with problems as they emerge. There may be dangers in this for consumers who might well end up paying the price for this early freedom. Regulatory uncertainty may also prove a greater brake on innovation than bad regulation.

The UK approach, outlined in a white paper in 2023, suggests a likely path somewhere between the two, lighter touch than the EU and sector-specific regulation. Use of AI is encouraged for financial services especially in areas like fraud. The FCA is likely to have regulatory oversight.

Representing opposing sides were Emma Erskine-Fox advocating for the motion and Ben Cooper speaking against it, both from TLT’s Future Law team.

An initial poll via the AFC app showed significant support for the motion, with 67.4% in favour and 32.6% against.

For the motion: The case for regulatory clarity

She noted the transformative potential of AI in improving lending decisions, fraud detection, and operational efficiency. However, she warned that without defined guardrails, these benefits could be outweighed by risks like bias, lack of accountability, and security vulnerabilities.

Key points from Emma’s argument included:

Responsible innovation through clear obligations: Emma argued that vague regulations deter innovation more than clear, robust ones. She stressed that businesses, uncertain of legal boundaries, often avoid high-reward use cases, fearing regulatory repercussions. She advocated for explicit delineation of accountability across AI’s lifecycle to streamline collaboration with vendors and protect lenders and consumers alike.

Building trust in AI. Public scepticism surrounding AI remains high, with words like “scary” and “worry” dominating public sentiment surveys. Emma contended that comprehensive regulation is crucial to operationalise ethical principles, ensuring consistent enforcement and boosting consumer confidence.

Consumer protection and vendor accountability: Emma highlighted gaps in transparency, especially when vendors withhold algorithmic details to protect intellectual property. Clear regulations could bridge this divide, ensuring lenders can meet data protection requirements while providing consumers with the information they need to make informed decisions.

Energy consumption and needs-led AI: She cautioned against tech-driven adoption without a clear focus on solving genuine problems, noting the environmental costs of AI. Emma argued that robust regulation would help prioritise AI’s deployment where it delivers the most value.

Against the motion: The case for innovation freedom

He stressed the sufficiency of existing regulations and called for global harmonisation rather than stricter UK-specific rules.

Key points from Ben’s argument included:

Maintaining the edge in fighting financial crime: Ben highlighted AI’s unparalleled effectiveness in combating financial crime, reducing false positives by 60% and significantly improving detection rates. He warned that excessive regulation could hinder the financial sector’s ability to stay ahead of organised crime, which operates without legal constraints or governance hurdles.

Sufficiency of current oversight: Citing a Bank of England report, Ben noted that 84% of financial firms already have accountable AI governance structures. He argued that additional regulation could duplicate efforts, creating unnecessary bureaucracy while existing fines and enforcement mechanisms already deter malpractice.

Avoiding a global patchwork problem: Ben expressed concern about inconsistent regulations across jurisdictions, which could create exploitable grey areas for bad actors. He advocated for a unified global framework rather than piecemeal national rules, suggesting that international cooperation is key to addressing AI risks comprehensively.

The risks of stifling innovation: Ben cautioned against adopting an overly restrictive UK regime, warning it could drive investment and innovation to less regulated jurisdictions. He highlighted the competitive advantage of a principles-based approach, aligning the UK with the US rather than the more prescriptive EU AI Act.

Shifting audience opinion

The debate’s conclusion saw a dramatic shift in audience opinion: 38.2% supported the motion, while 61.8% opposed it. This inversion highlighted the compelling arguments from the opposition, ultimately earning Ben Cooper a champagne victory.

The reversal in audience opinion suggests that Ben’s emphasis on the practical challenges of over-regulation resonated strongly with delegates. His framing of AI as a tool in a high-stakes “arms race” and his call for global consistency may have appealed to industry professionals wary of excessive oversight.

Additionally, his pragmatic focus on leveraging existing governance frameworks likely eased concerns about regulatory gaps highlighted by Emma.

Audience winner

Anne Jenkins captivated the audience and earned the coveted champagne prize for posing the best question at the event. Her inquiry, directed to the “pro corner,” delved into the delicate balance between specificity in regulation and the risk of stifling innovation.

Anne asked whether overly detailed regulations, such as the EU AI Act, might tie stakeholders in knots given the diverse applications and rapidly evolving nature of AI technologies. She championed principle-based legislation as a more adaptable approach, emphasising its capacity to keep pace with technological advancements.

Emma’s response acknowledged the tension but argued for a middle ground: addressing gaps in UK regulation, particularly around vendors who develop AI solutions. Anne’s question sparked a lively debate, with Emma highlighting the need for clearer, more robust obligations in the supply chain and Anne countering with a vendor’s perspective on accountability. The thought-provoking exchange underscored Anne’s ability to frame a nuanced issue that resonated with all sides, making her a well-deserved winner.

Conclusion

The debate underscored the complexity of managing AI’s potential and risks within the auto and equipment finance industry.

The near-inversion of polling results reflected the audience’s engagement with both perspectives, leaving no clear consensus but a deeper understanding of the challenges ahead.

As the UK navigates its regulatory approach, the industry will need to balance innovation with responsibility, informed by debates like this one that spotlight the multifaceted implications of AI regulation.

Watch the ‘AFC Soapbox Debate’ in full here.