OpenAI, Anthropic eye investor funds to settle AI lawsuits, FT reports (www.reuters.com)

OpenAI and Anthropic are reportedly exploring the use of investor funds and self-insurance structures to cover potential multibillion-dollar copyright claims tied to training large AI models, the Financial Times reported. The article says OpenAI has secured up to $300 million of emerging-AI risk coverage through Aon (a figure disputed by one source), but that insurance capacity is far short of potential exposure. Anthropic is partly using its own funds and faces a related authors’ suit for which a U.S. judge preliminarily approved a $1.5 billion settlement. Reuters could not independently verify the FT report and the companies did not comment. For the AI/ML community this signals a major shift in how model builders manage legal and financial risk. Limited insurance capacity and the prospect of multibillion-dollar payouts could push firms toward “captives” or ring‑fenced investor-backed reserves, tighter licensing and data‑provenance practices, greater use of synthetic or licensed datasets, and more conservative deployment strategies. It also raises governance questions about investor liability and capital allocation for legal contingency. Practically, expect increased emphasis on auditable training pipelines, provenance tooling, and contractual licensing to reduce downstream copyright exposure—and potential impacts on research openness, model cost structures, and the pace of commercial rollout.