As the world of generative artificial intelligence continues to advance, businesses are increasingly aware of the various risks associated with AI projects. These risks range from cybersecurity issues to potential copyright infringement, inaccurate or biased outputs, misinformation, and the inadvertent leaking of proprietary data. For business technology executives, these concerns can keep them up at night. However, amidst these apprehensions, insurance companies are recognizing a unique opportunity.
AI Insurance: A New Frontier
Drawing inspiration from the growth of cybersecurity insurance, which saw a surge in demand following major data breaches in recent years, insurance providers are now venturing into the realm of AI insurance. They are offering financial protection against the failure of AI models. This development is particularly welcomed by corporate technology leaders who see these policies as a means to address risk-management concerns voiced by board members, chief executives, and legal departments.
Niranjan Ramsunder, Chief Technology Officer and Head of Data Services at UST, a digital technology and information-technology services firm, highlights the emerging interest in AI insurance. “You will find more and more people starting to ask, ‘Who takes the risk? How do you fund it? And can you take care of some of the risk for us?'” he says. These questions are at the forefront as businesses grapple with the evolving landscape of generative AI.
While it’s still early days, analysts suggest there is a palpable appetite for AI insurance, especially for coverage related to financial losses stemming from AI and generative AI technologies. Even existing liability or cybersecurity policies could soon be tailored to accommodate generative AI. Although there isn’t a clear-cut example of generative AI causing data leakage leading to damages, the potential for such scenarios is enough to spur interest in insurance options.
Avivah Litan, a Gartner analyst focusing on AI trust, risk, and security, predicts, “I would bet that over fifty percent of large enterprises would buy some of these insurance policies if they come out, and they make sense.”
Pioneering efforts in AI insurance
Munich Re, a German reinsurer, ventured into the AI insurance space in 2018 by offering coverage for companies selling AI services. Michael Berger, head of Munich Re’s Insure AI product, explains that they also insure enterprises developing their own AI models by covering financial losses resulting from mistakes that a human wouldn’t have made. This approach provides a safety net for businesses navigating the uncharted waters of AI development.
Another player in the AI insurance arena is Armilla Assurance, a Toronto-based startup that launched this year. They offer a product warranty, backed by reinsurers including Swiss Re and Chaucer, guaranteeing that AI models will perform as promised by their sellers. This assurance is a step toward instilling confidence in businesses looking to integrate generative AI into their operations.
Vendor initiatives and legal safeguards
Recognizing the concerns businesses have about incorporating generative AI into their operations, tech giants like Microsoft, IBM, and Adobe are taking steps to mitigate risks. IBM, for instance, recently announced that its standard contractual intellectual property protections will extend to the generative AI models it develops. Similarly, Adobe allows businesses to purchase IP indemnification to protect against potential copyright issues related to generative-AI-created content on its Firefly platform.
Microsoft, in a noteworthy move, committed to defending and covering the legal costs of lawsuits arising from a customer’s use of its generative-AI-based Copilot tools, provided the customers use the built-in guardrails to filter out copyrighted content. This commitment has been a game-changer, as it has encouraged technical and product teams to adopt GitHub’s Copilot generative AI coding assistant, knowing that Microsoft is financially committed to supporting its customers in legal matters.
Learning from cyber insurance
The evolution of cyber insurance offers valuable lessons for the emerging AI insurance sector. Cyber insurers have increased scrutiny of policyholders’ security arrangements, resulting in more expensive policies and coverage denials, particularly during the pandemic. With a rise in costly cyberattacks, insurers have also raised premiums and reduced coverage. AI insurance policies could follow a similar trajectory as underwriting methods adapt and insurers grapple with the complexities of AI-related claims.
However, AI insurance faces unique challenges. The absence of historical data on AI model performance and usage in business makes it difficult for insurers to assess risk accurately. Additionally, generative AI models are evolving rapidly, requiring dynamic risk-assessment methods.
