The human authorship requirement has a new level of legal durability.
The Supreme Court declined to review the lower court ruling, leaving the human authorship requirement as the standing standard for copyright protection, per reporting from the International Intellectual Property Law Association. A cert denial isn’t a ruling on the merits, SCOTUS didn’t endorse the lower court’s reasoning or write new law. What it did was remove the near-term possibility of a court-driven change to the standard. The human authorship requirement stands until a case reaches the Court that it chooses to take, or until Congress acts. Neither track is moving quickly.
The same week, a federal judge is reviewing *Bartz v. Anthropic* for final approval. The reported terms: a $1.5 billion settlement, a requirement that Anthropic destroy pirated training files, and approximately $3,100 per copyrighted work, per Courthouse News reporting. Final approval hasn’t been granted. The per-work figure and the file destruction requirement are each significant, the per-work figure sets a pricing benchmark that other AI companies with similar training data exposure can model against, and the file destruction requirement goes beyond financial penalty to require affirmative remediation.
Evidence
Why these two events together matter more than either one alone.
The cert denial and the settlement aren’t legally connected. The copyright standard case and *Bartz v. Anthropic* arise from different legal theories and different facts. But they arrived in the same week, and what they establish together is clearer than either event in isolation: there’s no near-term judicial or negotiated path that creates a broad exception to the human authorship requirement for AI training data use. The Court isn’t taking the case that would change the standard. The largest pending settlement in AI copyright is moving toward resolution on terms that presume the human authorship requirement’s validity.
What the settlement benchmark means for other AI companies.
The $1.5 billion figure, and the approximately $3,100 per-work rate reported by Courthouse News, give AI companies with similar training data exposure a reference point for financial modeling. That doesn’t mean every AI company faces the same exposure, *Bartz v. Anthropic* involved specific claims about specific training data, and the settlement terms reflect the particulars of that case. But compliance and legal teams at AI companies that have not yet conducted a training data audit now have a settlement benchmark and a clarified legal standard to work against. The question isn’t whether exposure exists in the abstract. The question is what a particular company’s training data profile looks like against the standard that just got durably confirmed.
Analysis
The cert denial and the Anthropic settlement aren't legally connected, but they closed in the same week. AI companies that haven't conducted a training data audit now have both a durable legal standard (human authorship confirmed standing) and a financial benchmark ($1.5B settlement, ~$3,100 per work, per reporting) to model against. The analysis question is no longer whether exposure exists, it's how large.
What requires watching.
Final court approval of the Anthropic settlement is still pending. If the judge requests modifications to the terms, particularly around the file destruction requirement or the per-work payment structure, those modifications could adjust the benchmark. The Copyright Office’s testimony before the Senate, which reinforced the human authorship standard from the administrative track, suggests Congress is receiving consistent messaging about the standard’s durability, but no legislation to alter it has been introduced that would change the near-term picture.
Don’t expect SCOTUS to weigh in on AI copyright in the next term without a circuit split significant enough to compel its attention. The current trajectory, cert denial, pending settlement, Copyright Office legislative posture, points toward the human authorship standard remaining durable for the foreseeable horizon.