By Prathiba Krishna, AI and Ethics Lead at SAS UK & Ireland

Artificial intelligence (AI) is transforming the way businesses operate, make decisions, and interact with customers. In the banking sector, AI offers unparalleled opportunities, from automating processes to improving fraud detection. However, with such immense potential comes significant responsibility.

As AI becomes more deeply embedded in financial services, it is essential to guide its development carefully, ensuring that it not only drives innovation but does so in a way that is responsible and effective.

Robust governance and transparent practices

The rise of AI is inevitable, but how we manage its development will determine its impact. Businesses and governments must take a proactive approach to mitigating the risks associated with AI while still maximising its benefits.

This is especially true for the financial industry, where data integrity, security, and trust are paramount. The key lies in robust AI governance and a comprehensive framework which balances innovation with the ethical, legal, and societal implications of AI technologies.

AI governance needs to be built on three core pillars: transparency, accountability, and collaboration.

Transparency is crucial for building trust in AI systems and financial institutions must make AI processes visible and explainable to both customers and regulators.

Transparency also helps to demystify AI, allowing stakeholders to understand how decisions are made and at what stage - whether it’s detecting fraud or approving loans. By being open about how AI models work and evolve, banks can foster a culture of trust and accountability.

Accountability is also necessary to ensure that AI systems are developed and deployed responsibly. Financial institutions must be accountable for the outcomes of their AI models, especially when those models have the potential to impact people’s livelihoods.

Whether it’s an algorithm that wrongly flags a transaction as fraudulent or a bias in a credit scoring model, banks must have clear mechanisms to address any unintended consequences swiftly.

Continuous collaboration between businesses and regulatory bodies

Collaboration between financial services businesses, regulators, and governments, will be required to shape AI’s future in the banking industry. As AI evolves at an unprecedented pace, regulatory frameworks must adapt to the complexities of this technology.

Financial institutions must work closely with regulators to create policies that not only promote innovation but also ensure that AI is used safely. This requires a forward-thinking approach, where regulations can evolve in tandem with technological advancements while remaining flexible and adaptable.

One of the most important aspects of this is ensuring that AI regulation remains technology-neutral and responsive to the specific risks posed by AI in banking. Issues such as data privacy breaches, algorithmic biases, and cybersecurity threats must be carefully considered.

For instance, regulations should address how personal data is collected, processed, and discarded or deleted (when appropriate) in cloud-based AI-driven systems to maintain customer trust. At the same time, policies must ensure that AI systems minimise the impact of discriminatory biases and that robust security measures are in place to defend against cyberattacks.

Financial institutions must actively participate in these discussions, helping shape regulations that not only protect consumers but also encourage AI experimentation and innovation. By fostering open dialogue and collaboration, the banking industry can strike a balance between technological progress and ethical responsibility, ensuring AI’s benefits are fully realised.

Using AI to detect fraud and data biases

For institutions, leveraging AI’s potential to prevent harm while mitigating the associated risks will be critical to the banking industry’s success in the coming years - especially in terms of fraud detection.

AI’s ability to process large volumes of data in real time allows it to identify suspicious patterns and behaviours that traditional methods may miss. Machine learning models can spot unusual transaction trends or deviations in user behaviour, enabling financial institutions to catch fraud early and reduce losses.

AI also brings speed and precision to fraud detection, allowing institutions to react to potential threats in real time. For example, it can flag an unusually large transaction made from an unfamiliar location and temporarily freeze the account until further verification is completed - not only helping to prevent fraud, but also protecting customers from significant, potentially irreversible financial harm. This also protects institutions from having to reimburse victims of authorised push payment fraud up to £85,000 under new UK regulations that came into force from 7 October.

However, AI must be carefully managed to minimise the impact of any biases in the data models. If left unchecked, biased models can disproportionately target specific demographic groups, reinforcing negative stereotypes and leading to unfair treatment.

Financial institutions need to actively work to minimise these biases from AI models. This can be achieved through techniques like fairness-aware machine learning where algorithms are explicitly designed to account for and tackle bias.

The path to trustworthy AI development in the banking industry

The path to trustworthy AI development in banking isn’t without its challenges. But with robust governance, transparent practices, and a commitment to responsible innovation, AI can help financial institutions not only address current challenges but also excel in the future digital landscape.

To achieve this, embracing these principles is crucial. Banks can cultivate an environment that harnesses AI’s potential while considering customer welfare and societal impact. This means not only adhering to regulatory requirements but also going beyond compliance to establish best practices in trustworthy AI deployment.

As the landscape of financial services continues to evolve, the focus on trustworthy AI development will be paramount. The ability to navigate ethical dilemmas, foster trust, and create inclusive technologies will set forward-thinking banks apart in a competitive marketplace.

Ultimately, trustworthy AI is not just a regulatory obligation; it’s a strategic imperative that can drive long-term success and customer loyalty. By prioritising transparency, accountability, and collaboration, banks can unlock the transformative power of AI, ensuring a sustainable and customer-centric approach that safeguards the interests of their customers and communities.

Prathiba Krishna, AI and Ethics Lead at SAS UK & Ireland