Balancing Opportunity and Risk: Using Generative AI in Regulated Industries

For technologists and business strategists, artificial intelligence offers an intriguing, sometimes confounding duality. AI offers true game-changing opportunities to amplify the impact of organizational data sets. However, these opportunities must be measured against potential risks that – if not properly managed – could endanger a business and its customers. For organizations in regulated industries, such as credit unions, balancing business opportunity and risk with AI takes on special importance.

The great promise of AI, and new tools such as generative AI, is a democratization of sorts, with small- to mid-sized financial institutions beginning to enjoy the use of tools that were previously reserved for organizations that could afford to develop the technology. For example, in the United States, the vast majority of the more than 10,000 small- to mid-sized financial institutions that serve consumers use traditional methods to decide to whom to lend and provide other services. Using AI and machine learning technology to elevate the sophistication of their many products will result in more surgical data- and analytics-driven decisioning to ascertain creditworthy consumers. This would be a win-win for both lending institutions and consumers resulting in greater financial inclusion.

Unlocking the promise of generative AI requires a thoughtful and deliberate balance between innovation and opportunity with risk management. If a financial institution is too focused on the risks, it may end up in a standstill and underutilizing the technology’s potential. In this case, innovations from its competitors or other market players could leave it behind.

Conversely, a less cautious approach to using generative AI may result in analytics models that produce hallucinations (i.e., incorrect or misleading results), bias (i.e., results that reflect human biases, including social inequality), toxicity (i.e., a rude, disrespectful or unreasonable response) or other negative outcomes. In regulated industries such as banking and financial services, these outcomes could have significant consequences.

To avoid these pitfalls, here are three approaches to navigate the corporate use of generative AI:

1. Promote an active dialogue across risk, AI and tech functions. Given the fast-moving pace of generative AI, there is a need to evolve approaches on how to manage risks. Promoting an active dialogue across risk, AI and tech functions can help a lot with this. The natural operating rhythms between these organizations typically don’t allow for a degree of joint and collaborative innovation within their approaches. So, putting this in place can help.

2. Work backwards from tangible business impact. With something as new and shiny as generative AI, it can be easy to focus on strategic or operational improvements that will not realize value for the organization. Selecting use cases with rigorous prioritization and accountability on business impact (internal or with member) will help identify the risks that really need to be addressed to achieve the impact. One prime example of an application of generative AI to enhance internal operations is in member service. The use of this technology can better support answering questions from consumers and educating them on business-related topics or processes.

3. Invest in responsible AI. Organizations in regulated industries should invest the resources to ensure the proper governance of generative AI, especially as it relates to grassroots innovation that may touch consumer data. Guardrails to protect consumer privacy are vital. With emerging government regulation that carries moderate to severe penalties for misuse of AI systems and data, these organizations must design, develop and deploy AI systems that empower employee-based innovation, yet instill trust and confidence among partners and customers or members. Many forward-looking organizations have established AI risk councils that vet all applications of AI technology and accompanying use cases to manage risk and safeguard the responsible use of AI.

AI technology is a game-changer for organizations in regulated industries to accelerate and further strengthen their inclusion efforts, opening a world of opportunities for consumers with little to no credit. By applying AI and machine learning to previous risk models, they can generate new attributes that are suitable for thin/no file consumers. And that can have a transformative effect on their business and brand. However, the appropriate guardrails around internal AI-enabled grassroots innovation is necessary to minimize corporate risk and protect consumer data.

Shri Santhanam is an EVP and general manager of global analytics and AI at Experian.