Artificial Intelligence: Unlocking the Power of AI in Financial Services

innovative finance blog

Author: Christopher Foo, Innovate Finance Policy Associate

Artificial Intelligence (AI) is fueling both optimism and anxiety. Proponents of AI believe there is no going back for technology that makes life easier and more efficient. Opponents worry that AI poses risks to basic freedoms, security and even jobs. The financial services sector is keenly aware of the benefits and risks of AI and is seeking clear, contextual rules and guidance that enable continuous innovation without compromising service integrity.

The UK government’s Department for Science, Innovation and Technology (DSIT) is at the heart of AI regulation in the UK, and DSIT has released a white paper titled: An innovation-promoting approach to AI regulation. To ensure that the UK remains at the forefront of innovation and development, this white paper seeks to create a favorable regulatory environment to maximize the benefits of AI while managing AI risks through a clear framework of principles. We aim to build a

This blog provides a high-level overview of DSIT’s proposals and reviews the views expressed at the AI ​​Policy Roundtable at the 2023 Innovate Finance Global Summit (IFGS).

What is DSIT’s position on AI regulation?

AI is currently regulated through existing legal frameworks such as financial services and data protection regulations. However, gaps between existing regulatory mandates and conflicting and uncoordinated requirements from regulators risk unintentionally placing unnecessary burdens on businesses. This is against AI more generally, as regulatory gaps may leave risks unmitigated, which in turn may reduce public confidence in AI and delay AI adoption. and undesired effects.

DSIT recognizes this and is exploring a proportional, results-based approach to regulating AI. They argued that regulatory inconsistency can stifle innovation and competition, while companies have to sharpen their resources to comply with overly complex rules. We understand that your time may be spent innovating and developing new markets and opportunities. For any sector, including financial services.

What does DSIT propose?

An innovative principles-based approach to AI regulation is proposed by DSIT. This is consistent with Sir Patrick Vallance’s ‘The Rules of Innovation’ report. Review of Technical Regulations to Promote Innovation: Digital Technologies The report stated that a flexible regulatory approach would “strike the right balance between providing clarity and building public trust while also allowing for development, experimentation and deployment.” I’m here..

The DSIT framework is guided by six characteristics: Innovation Driven, Proportional, Reliable, Adaptable, Clear and Cooperative. These characteristics are underpinned by four key elements:

1. Defining AI based on its unique characteristics to support regulatory coordination.
2. Adopt a context-specific approach.
3. Provides a cross-cutting set of principles to guide regulatory responses to AI risks and opportunities.and
4. It provides new core capabilities to help regulators deliver AI regulatory frameworks, maximize the benefits of an iterative approach, and ensure framework consistency.

Essentially, DSIT does not assign rules or risk levels across sectors or across technologies. Instead, they propose regulating AI according to the results it may produce in a particular application. There are no blanket new rules that are future-proofed against unexpected new technologies.

However, DSIT recognizes the risks of such a decentralized regulatory framework. These risks include inconsistent application, enforcement and guidance among regulators. With this in mind, they propose a central cross-economic policy within government to identify, assess, prioritize and monitor cross-sectoral AI risks that may justify government intervention. We propose to create a risk function. For example, if a regulatory gap exists that justifies intervention and the intervention is outside the existing regulatory mandate, the Central Risk Function will identify actions that will be taken to address the gap. The white paper notes that regulators are given an opportunity to design this central risk framework.

How will the proposal be realized?

Existing regulators are expected to implement the framework and issue guidance for best practices through the five principles proposed as key elements of responsible AI design.

1. Safety, Security, Robustness.
2. Appropriate transparency and explainability.
3. fairness.
4. Accountability and Governance.and
5. Objections and Remedies.

These principles need to be applied proportionately to address the risks posed by AI within its mandate, in line with existing laws and regulations. Essentially, the above principles are complementary to existing regulations, providing clarity and less friction for companies operating beyond the confines of regulation.

It is worth noting that these principles are not initially statutory. This is to prevent legal requirements from stifling innovation and AI adoption. Instead, these principles will be introduced on a non-statutory basis and implemented by existing regulators. In practice, there is no consensus on the definition of AI, so regulators will use their domain expertise to tailor the implementation of the principles to the specific context of AI use. There is currently no timeline for enacting these principles, except that they may be enacted as needed after an “initial period of implementation” and “when time permits.”

DSIT believes that this principle-based approach, which has so far not been enshrined in law, can make the AI ​​regulatory framework agile and relevant. This light-touch approach is expected to satisfy AI stakeholders and advocates unfettered by restrictive laws while providing the guardrails necessary to guide responsible AI development.

What are some AI use cases?

Through stakeholder engagement and as expressed in several AI consultation responses (i.e. hyperlink: FCA seeks input on synthetic data ; UK Government Requests Evidence for “Establishment of Promoting Innovation Approaches to Regulate Artificial Intelligence (AI)” ; BoE/PRA/FCA Discussion Papers on “Artificial Intelligence and Machine Learning” (DP5/22 and DP22/4) ; and DSIT Consulting on “Promoting Innovation Approach to AI Regulation” ), there is a strong consensus that responsible use of AI could revolutionize financial services for the benefit of UK consumers and businesses. Members of Innovate Finance address key priorities in the financial services sector and share use cases demonstrating how companies can build and enable AI-enabled solutions to meet government and regulatory objectives Did. These include:

• Improve customer experience To:
  • Improved risk profiling capabilitiesThis could enhance suitability and affordability decisions for loans and consumer credit (improving issues related to financial inclusion and vulnerability).
  • Developing more bespoke pricing solutions By autonomous evaluation of datasets.
  • better explanation to customers Reasons for credit denial (including ways to improve your credit rating).and
  • Remote identity verification This eliminates the need to physically visit a bank branch and reduces the time it takes to open an account.
• Developing a dynamic model Its structure and parameters adapt to new data during deployment, providing an advantage over traditional static models that react slowly.
• Reduce operating costsmaking financial products more affordable and accessible.
• Facilitate rapid detection of fraud, financial crime, and other compliance issues., which can help reduce the pressure of procuring resources and reduce the backlog of anti-money laundering (AML) and other compliance procedures. Among other things, AI will enable more detailed verification of identity documents to ensure their validity.
• Strengthen investment strategy and Meet compliance and regulatory reporting Perform duties quickly with a high level of accuracy.
• Supporting the carbon dioxide reduction systemto help the financial services industry and the UK reach their net-zero targets.

What opinions were expressed at the round table?

It is worth noting that the AI ​​Policy Roundtable at IFGS 2023 was held shortly after the publication of the whitepaper, so the views may not have been fully formulated at that time. Nonetheless, the participants were encouraged by his DSIT UK AI Government Office staff who were present. This comes as officials reiterate their position that the government seeks to build public trust in AI by supporting innovation and getting regulation right, and the government wants to continue engaging with stakeholders. explains the strong interest in

Some (but not the only) key themes that emerged during the IFGS AI Policy Roundtable are:

• responsibility: Participants called for clarity on who is responsible for AI decisions and raised concerns about the costs of fintech companies being held responsible. On this basis, participants also discussed the best organization for determining, dividing and managing responsibilities.

• AI sandbox: While participants explored how AI sandboxes could drive innovation, DSIT officials noted that sandboxes go beyond simply testing products to determine the types of regulatory approaches that might be best suited for AI. He also pointed out that the purpose is to make a decision.

• Regulations: Participants debated whether AI regulation should be cross-cutting or sector-specific, as each area of ​​AI in financial services is inherently different.

• Interoperability: As many companies adopting AI operate in multiple markets, participants discussed the importance of ensuring that the UK’s approach to AI regulation is harmonized with other jurisdictions. .

• Reliability and explainability: Participants acknowledged that in order to create society’s trust in AI, companies need to be able to explain when they are adopting AI and how AI leads to decisions. This would require human intervention as a check and balance to AI decisions, which could make the technology even more reliable.

summary

A one-hour roundtable discussion is unlikely to cover all the issues discussed in relation to DSIT’s AI regulatory white paper. I can’t even do a brief (or perhaps quite long) blog post. However, stakeholders and Innovate Finance have had the opportunity to voice their views by answering the consultation questions in the white paper, which ended on June 21st. Governments and regulators need to continue to work with industry players to future-proof AI regulation and ensure subsequent steps, such as the introduction of AI regulatory principles and that future AI legislation is fit for purpose. there is. This is extremely important as new technologies emerge.

Innovate Finance will play a role in facilitating further discussion on AI regulation through a follow-up roundtable with DSIT officials from the UK Government Office for AI on 29 June 2023.

[ENDS]