Start with what isn’t happening: no AI copyright ruling from the US Copyright Office is imminent. No final rulemaking has been announced. If you’re waiting for the Copyright Office to resolve the question of whether using copyrighted works to train AI models is lawful before making product decisions, that wait isn’t ending this quarter.
That’s not a failure of the process. That’s what the process looks like.
Where the Copyright Office AI Work Actually Stands
The US Copyright Office’s ongoing consultation work on AI and copyright has been substantive. The office has published analysis on whether using copyrighted works to train AI models may constitute prima facie infringement of the reproduction right, a significant analytical step. That work covers training data composition, AI-generated works, and the boundaries of fair use in model development.
Skadden’s analysis of the Copyright Office’s published work characterizes the pre-publication document as weighing in on AI training and fair use without resolving it. That distinction is important. The Copyright Office has engaged seriously with the question. It hasn’t answered it with binding authority.
The consultation process is ongoing and advancing. But “advancing” and “concluding” are different things.
What This Means for Developers Using Copyrighted Training Data
Three things are true simultaneously, and all three matter for planning:
First, copyright liability for AI training data is a live legal question in the United States. Multiple cases are in litigation. The Copyright Office’s analysis acknowledges the reproduction right question is real.
Second, no US court or agency has issued a definitive ruling. Organizations building on or with copyrighted training data are operating in unresolved legal territory, not in a space where one side has clearly won.
Third, the Copyright Office’s process will eventually produce policy guidance. What form that takes, rulemaking, additional guidance documents, or legislative recommendations to Congress, hasn’t been determined.
What to Watch
The Copyright Office’s published work is the most authoritative reference currently available. Practitioners should track: any announcement of a final rulemaking process, publication of additional Copyright Office guidance documents, and developments in AI copyright litigation that may produce court decisions before the Copyright Office process concludes.
This brief is part of continuing hub coverage on US AI copyright policy. For the cross-border context and comparison of how the EU and UK are approaching similar questions, see the Three Governments, Three AI Copyright Approaches brief. For the Congressional response to the White House framework’s copyright position, see the earlier coverage here.
TJS Synthesis
The Copyright Office consultation process is a signal in itself. The pace of administrative IP rulemaking is measured in years, not months. Developers who need clarity on training data copyright liability are more likely to get it first from litigation outcomes than from a completed Copyright Office rulemaking. Building legal strategy around “the Copyright Office will resolve this soon” is a planning assumption that doesn’t hold up against the documented timeline of the process.