Reuters: OpenAI has asked a federal judge to throw out a lawsuit claiming ChatGPT crossed the line from software into unlicensed legal practice.
Nippon Life says a former employee used ChatGPT to produce a flood of court filings after settling a disability-benefits dispute, and argues OpenAI helped drive filings with no real legal purpose.
OpenAI’s defense: ChatGPT is not a person, has no law license, and does not “practice law” just because a user asks it to draft arguments, motions, or research notes.
The deeper issue is where responsibility sits when an AI helps a self-represented person create legal documents that may be weak, repetitive, or procedurally wrong.
OpenAI says the user made the choices, the judge had authority over the filings, and frustration with defending a messy pro se case cannot turn a toolmaker into a lawyer.
Courts are seeing more people use generative AI to write filings, which can lower access barriers while also increasing junk motions, fake citations, and procedural confusion.
---
reuters .com/legal/transactional/openai-says-chatgpt-is-not-lawyer-asks-court-toss-insurers-lawsuit-2026-05-18/
