对“AI惹祸”投保？保险公司“不敢接”

Core Insights - The insurance industry is becoming increasingly cautious about the risks associated with artificial intelligence (AI), leading to significant changes in policy coverage [1][2] - Major insurance companies are seeking to exclude AI-related risks from standard business policies due to concerns over the opaque decision-making processes of AI models [1][2] - Real-world incidents of AI-related claims are prompting insurers to act, highlighting the potential for systemic risks that could arise from AI failures [1][3] Group 1: Insurance Industry Response - Major insurers like AIG, Great American, and WR Berkley are applying to regulators to include exclusion clauses in their policies that specifically address liabilities arising from the use of AI technologies [1][2] - The shift in attitude reflects a growing concern that AI models can lead to numerous interconnected claims, creating unmanageable systemic risks for the insurance sector [2][3] - Insurers are particularly wary of the potential for a single AI model's failure to result in thousands of claims, which could overwhelm their capacity to pay [2] Group 2: Specific Incidents and Examples - Notable cases, such as a Canadian airline's chatbot generating false discounts and Google facing a $110 million lawsuit for erroneous AI search results, underscore the tangible risks associated with AI [1][3] - The engineering firm Arup lost $25 million due to fraud involving a digital clone of an executive, further illustrating the vulnerabilities that insurers are now hesitant to cover [3] Group 3: Limited Coverage Options - Some insurers are exploring limited coverage options, but these often come with strict limitations, such as QBE's policy capping AI-related fines at 2.5% of the total coverage [4] - Chubb has agreed to cover certain AI risks but has explicitly excluded broad AI events that could affect multiple clients simultaneously [4] - Legal experts warn that as AI-driven losses increase, insurers may begin to contest claims in court, potentially requiring a significant systemic event to prompt a change in their approach [4]