A registration gap can erase an author's settlement award

The writer's problem

Once the Anthropic settlement made copyright registration status the line between compensation and exclusion, many authors discovered for the first time that a publisher had never registered a book on their behalf. The Authors Guild published survey results on 20 August 2026 describing what members reported after the settlement's eligibility rules became public.

What the documents show

The guild's own release states this is a self-selected survey of its members, not a random or independently audited sample, and it says so explicitly. Of the books for which members reported a contractual registration requirement, the release states 93 percent of respondents said the publisher had not complied, a figure the guild itself immediately qualifies as reflecting who chose to respond rather than an industry-wide noncompliance rate. The release also states 321 of the 804 reported non-compliant books were published between 2019 and 2025, which the guild frames as evidence the problem is recent and ongoing rather than a legacy issue, and separately notes that timely registration is the specific eligibility requirement under the Anthropic settlement described in the guild's own settlement guidance. The release is explicit that a survey response alone does not establish a breach of contract, which would require examining each publishing agreement's actual terms.

The editorial choice

The guild's own recommendation, labeled here as editorial advocacy rather than fact, is that authors and agents should insist on enforceable registration deadlines in future contracts and should independently check the Copyright Office's public catalog rather than assume a publisher registered a book as promised. That advice follows from the survey's findings but goes beyond what the survey itself proves about any individual contract.

What stays with the author

Confirming whether a specific book was actually registered, and by when, remains a task only the author or their agent can verify against the Copyright Office's own records; the survey describes a pattern, not a verdict on any one title. The sources do not resolve how many of the reported failures affected books that would otherwise have qualified for the Anthropic settlement specifically, as opposed to registration gaps generally.

• Is a cited failure rate drawn from a self-selected survey, and does the source itself flag that limitation?
• Has the author independently checked the Copyright Office's public catalog for their own titles?
• Does the contract actually specify a registration deadline and assign responsibility for meeting it?

A high self-reported failure rate is evidence of a real and current problem, but it is not, on its own, a measurement of how often publishers as a whole meet their registration obligations.