Who covers AI business blunders? Some insurers cautiously step up

As businesses increasingly rely on artificial intelligence (AI) to manage operations and drive revenue, the insurance industry is beginning to adapt to the unique risks associated with these technologies. Some insurance firms are stepping up to provide coverage for potential AI-related errors, while others remain hesitant, reflecting a cautious approach to this evolving landscape. This shift is underscored by the emergence of 'agentic AI,' where AI systems operate autonomously, minimizing human oversight and intervention.

Phil Dawson, head of AI policy and partnerships at the specialist insurer Armilla, highlights the intent behind adopting advanced AI technologies: to replace human assistance in decision-making processes. However, this shift poses significant challenges to traditional insurance models, as the reliance on AI introduces new types of liabilities. Dawson notes that while companies strive to refine AI technologies, the risk of errors—such as 'hallucinations,' where AI confidently generates inaccurate or fabricated outputs—remains a pressing concern.

The concept of 'silent coverage' has been prevalent in the insurance sector, wherein AI-related liability risks have been implicitly included in existing policies. Analysts Sonal Madhok and law professor Anat Lior, in a research paper from the brokerage firm Willis Towers Watson, argue that the current scenario mirrors the early days of cybercrime, where liability coverage was not explicitly defined. They predict that insurance policies will soon evolve to explicitly address AI-related risks, marking the end of the silent coverage era.

Jonathan Mitchell, head of the financial sector practice at brokerage firm Founder Shield, observes that insurers are moving away from a passive 'wait-and-see' strategy regarding AI mishaps. Some standard insurance policies now incorporate 'absolute AI exclusion' clauses, which specifically deny coverage for incidents arising from AI errors. This shift indicates a growing recognition of the unique challenges posed by AI technologies.

Dawson provides an example of a commercial real estate firm that sought to insure its AI agent as if it were a regular employee but ultimately had to revert to a specialized policy. This highlights the complexities of insuring AI systems and the need for tailored coverage solutions.

In response to the evolving landscape, Founder Shield has developed policies that specifically address 'AI malfunction and hallucination' scenarios, extending coverage beyond traditional computer networks to include real-world impacts, such as erroneous inventory orders. Armilla takes a proactive approach by testing AI models for vulnerabilities before offering coverage and ensuring that clients' risk management frameworks comply with international standards. However, like other insurers, Armilla retains the right to decline coverage for certain high-risk areas, such as medical diagnostics and mental health applications.

Munich Re, a leading global insurer, offers coverage for companies involved in both designing and utilizing AI models. Michael von Gablenz, head of AI insurance at Munich Re, acknowledges that the risk of AI systems making errors cannot be entirely mitigated due to their inherent statistical nature. He emphasizes that all statistical models carry a degree of uncertainty, which complicates the insurance landscape.

Despite these challenges, the emergence of AI-related risks presents significant opportunities for insurers. Von Gablenz estimates that the market for AI insurance could surpass that of cybersecurity insurance, reflecting the growing importance of AI in various sectors. The Deloitte Center for Financial Services projects that the global AI insurance premium market could reach as much as $4.8 billion by 2032, indicating a robust future for this niche within the insurance industry.

In conclusion, as businesses continue to integrate AI technologies into their operations, the insurance industry is gradually adapting to the associated risks. While some insurers are beginning to offer tailored coverage for AI-related mishaps, others remain cautious, highlighting the need for further clarity and understanding of the liabilities involved. As the market for AI insurance develops, it will be crucial for both creators and technologists to stay informed about these evolving policies and the implications for their work.

Why it matters

The rise of AI technologies in business operations necessitates a reevaluation of traditional insurance models. As AI systems become more autonomous, the potential for errors increases, prompting insurers to develop new policies that explicitly address these risks. This evolution in insurance coverage is vital for businesses that rely on AI, as it provides a framework for managing potential liabilities and protecting against unforeseen consequences.

Moreover, the anticipated growth of the AI insurance market signals a burgeoning opportunity for both insurers and businesses alike. As the landscape of AI continues to evolve, understanding these insurance dynamics will be essential for creators and technologists seeking to navigate the complexities of AI implementation and risk management.