AI Act: Europe’s bid for trustworthy AI development and global leadership – Earnix

The European Union AI Act (Regulation (EU) 2024/1689) officially entered into force on August 1, 2024, and will be fully applicable by 2026, with provisions rolling out in the coming months. As the first comprehensive legal framework for artificial intelligence (AI) globally, it marks a pivotal moment in shaping the future of AI development and regulation.

In tandem with this, the European Commission launched the AI Pact to encourage early compliance with the Act’s obligations. The Pact aims to foster trustworthy AI in Europe by addressing potential risks, ensuring safety, and safeguarding fundamental rights.

The Act sets clear obligations for AI developers and deployers, particularly in high-risk AI applications, while aiming to reduce administrative burdens, especially for small and medium-sized enterprises (SMEs).

The AI Act categorises AI systems into four risk levels: unacceptable, high, limited, and minimal/no risk. High-risk systems, such as those used in critical infrastructure, law enforcement, and education, face stringent requirements, including risk assessments, robust datasets, traceability, human oversight, and security measures. Remote biometric identification for law enforcement is largely prohibited, with narrow exceptions.

For limited-risk AI, such as chatbots or AI-generated content, transparency obligations are introduced, ensuring users are informed when interacting with AI systems. Minimal-risk AI, like video games or spam filters, can be freely used.

Additionally, the Act establishes transparency and risk management measures for general-purpose AI models and strives to be adaptable to future technological advancements. The European AI Office will conduct enforcement, working with member states and international stakeholders.

As AI-driven technologies become integral to insurance, these regulations and others such as the Digital Operational Resilience Act (DORA) will challenge the industry to ensure compliance while continuing to innovate.

For many insurers, the transition to AI is already underway, with more than two-thirds of respondents from across the insurance landscape to the recent Earnix Trends survey expecting to deploy AI models that make predictions based on real-time data in the next two years. As AI and analytics become central to customer service, risk assessment, and decision-making, financial institutions must integrate these technologies while navigating complex regulatory requirements.

Transparency is key

AI systems must be transparent, free from bias, and easily explainable. These criteria are not only essential for regulatory compliance but will also help build trust with consumers and regulators. For reinsurers, balancing innovation with responsibility is crucial to ensuring that AI’s benefits do not come at the expense of fairness or accountability.

The AI Act is particularly relevant in this context, mandating greater transparency in AI systems, especially for models used in risk assessments or underwriting decisions. This is a pivotal moment for reinsurers who are increasingly leveraging AI for critical tasks such as loss reserving and claims management. By analysing vast amounts of historical claims data and real-time market trends, AI-powered models improve reserve estimation, leading to better capital allocation and risk management.

AI is also transforming pricing strategies. With AI-driven pricing engines, insurers can create more granular pricing models that consider a wider range of variables, including real-time market conditions. As AI evolves, its role in regulatory compliance will become even more important. With regulatory fines a growing concern, it is no surprise that 70% of insurance executives report prioritising compliance, particularly in light of new laws like the AI Act.

Beyond pricing and risk management, AI is helping insurers reduce claims leakage-where millions of dollars are lost each year due to missed or inefficiently processed claims. By enhancing claims processing, insurers can mitigate operational inefficiencies and protect their bottom line. The next frontier for AI in reinsurance lies in generative AI (GenAI), already being integrated into workflows such as underwriting and exposure management.

Legacy integration does not have to be a hurdle

Legacy integration doesn’t have to be a hurdle—with the right partner, it can be a strategic advantage. While integrating AI into legacy systems presents complexities, the right approach can streamline the process, optimising both technology and organisational workflows.

With careful planning and investment, AI can enhance accuracy, efficiency, and decision-making in claims management and underwriting. Proactively addressing data integrity and ethical considerations ensures compliance and builds trust, turning potential challenges into opportunities for innovation and growth.

One promising trend is the rise of the Chief AI Officer (CAIO), a key role emerging to address these challenges. The CAIO will help organisations navigate regulatory complexities, close skills gaps, and maintain a competitive edge by ensuring responsible AI deployment, while reporting AI APIs at board level.

Beyond operational benefits, AI should prove an indispensable additional tool in the exposure management toolkit when it comes to addressing climate risks - Capgemini recently reported that more carriers are expected to update their models to incorporate climate related risks and appropriate pricing, and are improving climate-related property risk evaluations using predictive analytics models to bolster risk management and informed underwriting.

With AI’s ability to model complex scenarios, such as rising sea levels and extreme weather events, insurers can better assess risks that are not represented in our existing historic data sets, and identify protection gaps. AI-driven scenario modeling is now enabling insurers to address emerging risks, like those related to climate change.

In high-risk areas, AI could increasingly help insurers better understand and manage the growing challenges posed by extreme weather events. This is vital as insurers seek to bridge the protection gaps exacerbated by climate risks. Collaboration with regulators, climate scientists, and policymakers is essential to ensure that AI-driven solutions are both equitable and actionable - striking the balance between unlocking new opportunities while maintaining the highest standards of transparency and fairness.

By Robin Gilthorpe, CEO, Earnix