Generative AI is steadily moving from an emerging technology to a strategic tool within financial institutions. According to Anthony Raba, senior client advocate, corporate risk and brokerage at WTW, most firms are in the early adoption phase, focusing on establishing governance frameworks, assessing risks and understanding how to integrate generative AI into their broader plans.

Initially, AI is being used in administrative processes to enhance efficiency by increasing productivity and automating routine tasks. Over time, it is expected to play a more prominent role in the front office, helping to engage clients and drive the development of innovative advisory and financial products.

“Managing the risks associated with the new generation of AI is like solving a complex puzzle – there is no single insurance policy that covers all potential risks. Instead, you must assemble a comprehensive risk management strategy using different policies from your insurance portfolio,” said Raba.

As a general-purpose technology, AI can be integrated into a variety of applications, impacting nearly every aspect of a financial institution’s operations. This integration can significantly alter existing risk profiles and introduce new risks that may not otherwise be addressed.

These risks include bias and interpretability, as the complexity of AI models makes it difficult to interpret their decisions, and these systems may inadvertently reflect and perpetuate biases present in their training data.

Concerns about quality and hallucination are also prevalent, as a lack of training data or lack of originality could lead to the production of plagiarized content. Over-reliance on AI poses another risk, as increasing capabilities could lead to over-reliance without proper critical evaluation.

Companies that rely on third parties may face risks to the reliability, security, and continuity of these services, while internal resources require investments in talent and data storage. Rapid developments in AI may outpace regulatory developments, leaving companies in uncertain legal environments.

The complex algorithms used by AI and the use of data increase risks related to intellectual property and data privacy. AI tools can be misused for social engineering attacks, financial fraud, and the creation of complex fake identities. Identifying errors when AI makes decisions can be difficult. The unpredictability of the resulting output complicates the process of establishing clear guidelines for content moderation.

Biased or flawed data can lead to inaccurate results, amplifying miscalculations and inequalities. AI applications can also raise ethical issues, such as in facial recognition, surveillance, or autonomous decision-making in sensitive areas such as healthcare and criminal justice. Finally, the adoption of AI can lead to job displacement and unemployment.

Managing the risks associated with the new generation of artificial intelligence

To effectively manage these risks, financial institutions need to consider a multi-pronged insurance approach. Cyber ​​insurance is critical to addressing the growing vulnerabilities and data privacy concerns created by the AI ​​generation.

According to the 2024 Global Executive and Director Survey Report, cybersecurity remains a top priority for business leaders. Liability insurance for hiring practices will be essential as the AI ​​generation drives significant changes in the workforce, necessitating new approaches to talent acquisition and team structuring. This shift will require companies to adapt hiring practices and reorganize teams to fully leverage the capabilities of the AI ​​generation.

Professional liability/errors and omissions (PI/E&O) insurance covers the use of AI in professional services, including financial advisory and portfolio management. Casualty insurance, often associated with “slip and fall” accidents, can also cover personal and advertising injuries, which is important when AI-generated content inadvertently uses copyrighted material. Intellectual property insurance is increasingly required by financial institutions to protect against the infringement risks posed by AI.

Property insurance, a first-party coverage, protects physical infrastructure and business continuity, although it does not cover many of the regulatory and liability risks posed by first-generation AI. Fidelity/bond/crime insurance protects against losses from employee theft, fraud, and computer crime, all of which are impacted by the adoption of first-generation AI.

Securing directors and officers is essential to protecting leadership during the adoption phase of next-generation AI, and addressing the various risks and responsibilities that flow to company leadership.

“Managing AI-related risks requires a comprehensive, integrated approach. Financial institutions need to consider how each part of their insurance coverage can be integrated together to form a complete risk management strategy. Addressing the unique challenges posed by AI across different policies helps protect against potential exposures,” said Raba.

This strategy should include scenario analysis and crisis response planning to uncover coverage gaps and necessary improvements. Policy coordination is vital, ensuring consistency across policies by scrutinizing coverage provisions and other items. Advanced analytics should be leveraged to understand a company’s unique risk profile rather than relying solely on benchmarking.

Proactive engagement with insurers can demonstrate a company’s superior governance framework and understanding of risk, which can be critical to renewal success as the market and risks continue to evolve.

What do you think of this story? Feel free to share your comments below.