Ninth Circuit affirms dismissal of programmers' DMCA §1202(b) claims over GitHub Copilot and OpenAI Codex outputs

What happened

On 16 September 2026, the U.S. Court of Appeals for the Ninth Circuit issued a published opinion in Doe v. GitHub, Inc., No. 24-7700, affirming dismissal of the plaintiff programmers' claims under DMCA Section 1202(b) against GitHub, Microsoft, and OpenAI. The programmers alleged that Copilot and Codex, trained on public repositories including their open-source code, generate code without the attribution, copyright notices, and license terms (copyright management information) attached to the originals. The court held that generating a new work that never carried CMI is not 'removing or altering' CMI from an existing work, while noting that near-verbatim reproduction could be circumstantial evidence of removal. The panel declined to consider the plaintiffs' 'input' theory — that CMI was stripped before training — finding it forfeited, and expressed no view on ordinary copyright infringement by the outputs.

Why it may matter to you

It is the first appellate ruling on the DMCA theory programmers used to challenge AI coding tools built on open-source code. The output-side §1202(b) route is now closed in the Ninth Circuit; the training-stage question was not decided.

What is not yet known

Whether the plaintiffs seek rehearing en banc or Supreme Court review; earlier secondary reports conflicted on panel versus en banc posture, and the opinion PDF itself was not captured this run.