AI As a Tool for Inclusion, Not Exclusion

This piece is by Erez Barak, CTO at Earnix

The role of Artificial Intelligence (AI) in the UK insurance sector has sparked widespread debate, particularly with the recent concerns raised by the Financial Conduct Authority (FCA). In a 2024 warning, the FCA cautioned that AI-driven hyper-personalization could make some individuals “uninsurable,” raising concerns that technology could exacerbate inequalities in the market. While this perspective highlights potential risks, it should also consider AI’s transformative power to drive inclusion rather than exclusion, especially when we look at “agentic” AI – systems capable of autonomous actions and decision-making that can enhance fairness and access across the sector.

What Did the FCA Say About AI and Inclusion?

In the autumn of 2024, the FCA voiced concerns that AI might render certain consumers uninsurable. Their fear stems from the risk that AI-enabled hyper-personalisation, where policies are tailored and priced based on individual customer data, could make it more difficult for certain people, particularly those with pre-existing health conditions or limited access to technology, to obtain coverage. The use of real-time data to adjust risk assessments and premiums could inadvertently exclude individuals from the market altogether. The regulator has called for the industry to balance innovation with consumer protection and has pledged to monitor AI’s integration within the insurance sector closely.

Why Does the FCA Believe This?

The FCA’s concerns are rooted in the potential for AI to perpetuate or even amplify existing biases in underwriting practices. In a sector where risk data has historically been skewed, there’s a fear that AI could lock certain individuals out of coverage by relying on data that fails to account for the full diversity of the population. For instance, people with complex medical histories, non-traditional job roles, or those living in underserved regions might be disproportionately affected by automated decision-making systems. Additionally, AI systems, if improperly designed, could perpetuate inequalities present in historical data, resulting in higher premiums or even policy denials.

The Opposing View: AI’s Potential to Drive Inclusivity

However, these concerns overlook a powerful counter-narrative: AI, particularly “agentic” AI, has the potential to increase inclusivity in the insurance industry. Agentic AI refers to systems that are not just tools for data processing but are capable of autonomous actions and decision-making. Such systems, when designed ethically, can improve fairness by taking action based on a more accurate understanding of individual needs and circumstances.

Agentic AI can enable insurers to move away from outdated, broad-brush underwriting methods. By incorporating a wider range of data points – including real-time health monitoring, lifestyle choices, and more nuanced demographic factors – agentic AI can offer products specifically tailored to those who are often overlooked, such as gig economy workers or individuals with unique medical histories. This level of autonomy allows AI to make decisions that better reflect the diversity of the population, enabling fairer pricing and improving access for those who have been traditionally excluded by legacy systems.

AI’s ability to refine flawed historical data is another crucial factor in reducing bias. Traditional underwriting models often rely on outdated or incomplete data, which can lead to mispricing and exclusion of certain groups. Agentic AI, in contrast, can process vast amounts of diverse data to identify previously unrecognised risk factors, allowing insurers to offer more accurate pricing and improved coverage options. This shift could not only lead to fairer pricing but also expand access to those previously marginalised or underserved.

Moreover, Agentic AI opens up new opportunities for global inclusivity. Emerging markets, where risk data is often sparse, stand to benefit from AI’s capacity to process limited information and generate accurate risk models. In regions where traditional data sources are lacking or unreliable, AI can facilitate coverage by identifying and quantifying risks more effectively. This can bring affordable insurance to populations in developing economies, where access to financial services is often limited.

When used responsibly, Agentic AI has the potential to drive fairness and transparency in the insurance industry, counteracting the discriminatory practices of the past. By continuously improving and refining AI models, insurers can ensure that risk assessments reflect a broader and more accurate picture of individual circumstances, reducing the likelihood of exclusion.

Conclusion

In the rapidly evolving insurance industry, it is essential to recognise that various innovative ideas are not in competition but rather complement each other to drive overall growth. By integrating advanced technologies such as artificial intelligence (AI) and machine learning, insurers can enhance their decision-making processes, improve customer experiences, and streamline operations.

For instance, AI-powered analytics can provide deeper insights into customer behavior, enabling insurers to tailor their products and services more effectively. At the same time, machine learning algorithms can automate routine tasks, freeing up valuable resources for more strategic initiatives. Together, these technologies create a more efficient and responsive insurance ecosystem, ultimately benefiting both providers and policyholders.

Moreover, collaboration between different stakeholders, including technology providers, insurers, and regulatory bodies, is crucial for fostering innovation and ensuring that the industry adapts to changing market dynamics. By working together, these entities can develop solutions that address emerging challenges and capitalize on new opportunities, driving sustainable growth for the entire sector.

The debate surrounding AI in insurance is complex, and the FCA’s concerns are not unfounded. However, it is crucial to recognise that AI, particularly Agentic AI, when implemented ethically, with best practices in mind, and with a focus on inclusivity, can serve as a powerful tool for positive change. Far from increasing exclusion, AI has the potential to foster greater fairness, affordability, and access to insurance products for all individuals – regardless of their health, job, or background. This is a vision of AI in insurance that embraces diversity, reduces bias, and ultimately works to lower barriers to accessing insurance as a tool for resilience for all.