▲ OpenAI (File photo)
Arguments have been raised that if artificial intelligence (AI) goes out of corporate control and causes an accident, legal responsibility could even extend to top executives such as OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei.
The Financial Times (FT) reported on the 6th local time that the insurance and legal industries are preparing for the possibility of large-scale damage lawsuits that AI could trigger.
With growing awareness of AI risks recently—such as OpenAI's AI agent hacking Hugging Face—discussions surrounding the scope of liability are gaining momentum.
An analysis of more than 300 AI-related disputes by insurance broker Aon showed that insurers could be required to pay out claims across various products, including crime insurance, intellectual property insurance, cybersecurity insurance, and technology errors and omissions insurance.
In particular, arguments emerged that the executives of AI companies cannot be entirely free from liability for damages.
Regarding the Hugging Face hacking incident, Tim Rayner, head of claims in the UK at insurance information firm Verisk, argued, "Ultimately, OpenAI's CEO should be held responsible," adding that it was due to a "lack of control over the business."
He went on to say, "Every CEO has a responsibility to ensure that their company is properly managed and controlled," and "AI does not change that."
The insurance industry is paying attention to the Hugging Face incident because executives could become targets of lawsuits in the future based on accidents involving uncontrolled AI agents.
Among insurance products, there is "directors and officers (D&O) liability insurance," which covers losses if corporate executives are sued over management decisions or statements.
This means that if OpenAI is subscribed to this insurance, it could claim payouts for losses arising from lawsuits targeting CEO Altman.
However, there are still no legal precedents regarding holding developer executives accountable for the actions of AI.
Aaron LeMarquer, head of insurance disputes at law firm Stewards, said, "If OpenAI shareholders prove that management's failure in risk management caused losses to the company, there is also the possibility that they could file lawsuits against the directors."
LeMarquer projected that future damage lawsuits against AI companies could evolve in a similar fashion to past environmental pollution lawsuits, or lawsuits against tobacco and pharmaceutical companies.
This implies that large-scale lawsuits targeting AI companies themselves could be filed.
The insurance industry believes that lawsuits could be brought against AI developers across various sectors, including product liability and privacy violations.
In fact, Anthropic agreed in July to pay 1.5 billion dollars (approx. 2 trillion KRW) in connection with a class-action lawsuit filed by authors alleging unauthorized use of copyrighted works during the training process of its AI model.
It was also reported that Anthropic plans to include in its future initial public offering (IPO) prospectus that its technology "could pose an existential risk to humanity."
※
Copying, redistribution, and unauthorized use in AI training are strictly prohibited.