A post argues AI outputs and chain-of-thought are not copyrightable in the US
kevinsxu · x · 2026-07-23
AI model outputs and chain-of-thought are not copyrightable in the US, says an X post
The post argues that under current US law, an AI model’s output and chain-of-thought are not copyrighted and are not treated as trade secrets, so calling distillation “IP theft” is, in the author’s view, a weak legal argument.
It frames the issue as a matter of law rather than intuition:
- laws apply until they are changed
- public feelings about ownership do not determine legal status
- the US legal system is presented as self-correcting and less swayed by lobbying over time
The post also uses a patent example to make the point that legal categories can be surprising, but still real until the law changes.
More from Safety
- Meta joins frontier labs reporting CoT controllability evals in Muse Spark 1.3 card — tomekkorbak · 2026-10-27
- Meta TBD underwent pre-launch red-teaming by 5 third-party safety firms — maksym_andr · 2026-10-27
- Is ChatGPT showing your chats to others? Screenshot sparks concern — chrisalbon · 2026-10-27
- Do canary tokens still work in the coding-agent era? Fake AWS keys may lure agents — kuza55 · 2026-10-27
- LLM bias judges name groups instead of saying 'unknown', refusing only 0.4% of the time — ZhijingJin · 2026-10-27
- ChatGPT will render genitals if you ask it to generate an X-ray of yourself — flowersslop · 2026-10-27