Authors and publishers who maintain clear ownership records, timely copyright registrations, and basic technical controls are better positioned to claim their share of AI copyright settlement funds. This practical guide shows the lasting steps that apply to the current Anthropic settlement and to future cases.
What You Need
- A complete inventory of your published works (exact titles, ISBNs or ASINs, publication dates, publishers, and any existing registration numbers)
- An account on the U.S. Copyright Office Electronic Copyright Office (eCO) system
- Digital deposit copies of each work in an accepted file format
- Payment method for the registration fee (credit/debit card or ACH)
- Publishing contracts or other proof of ownership and rights splits
- Access to the official settlement Works List lookup for any active case
- Control over your website so you can edit robots.txt
Step-by-Step Guide
Step 1: Build a complete works inventory
Create a spreadsheet listing every title you own or co-own. Record the exact title, author names, ISBN or ASIN, first publication date, publisher, and any copyright registration number you already have.
Note co-authors, exclusive licenses, and whether the work was made for hire. Keep digital copies of copyright pages, contracts, and cover images in a secure folder. Update the inventory whenever a new work is published.
Step 2: Search existing registrations in the Copyright Public Records System
Before filing anything new, search the official Copyright Public Records System (CPRS) to see whether a work is already registered. The CPRS replaced the older public catalog and holds registration and recordation records from 1978 onward.
Go to the CPRS site and search by title, author name, ISBN, or registration number. Record every matching registration number and effective date in your inventory. Works that lack timely registration are usually ineligible for settlement class membership.
https://publicrecords.copyright.gov
Step 3: Register unregistered works through eCO
Registration is done online through the Electronic Copyright Office (eCO) system. Create an account if you do not already have one, then start a new claim. Choose the Single Application only when one individual author created one work, owns all rights, and the work is not made for hire; otherwise use the Standard Application.
Complete the application accurately, pay the fee, and upload the deposit copy of the work. The effective date of registration is the day the Office receives a complete submission in acceptable form. For Anthropic-type settlements, eligibility typically requires registration within five years of publication and either before the infringing download or within three months of publication.
- Single Application fee (current): $45 for qualifying single-author works
- Standard Application fee (current online): $65
- Processing times vary; most online applications receive a certificate within a few months
https://www.copyright.gov/registration/
If the work contains more than a minimal amount of AI-generated material, disclose that material and exclude it from the claim in the Limitation of Claim section. Claim only the human-authored portions. Purely AI-generated material is not registrable.
Step 4: Search active settlement Works Lists
Use the official lookup tool for any open settlement. For the Anthropic case, search by ISBN, ASIN, title, author, or registration number. Confirm which of your works appear on the Works List and note the listed rights holders.
Record the claim status of every match. If a qualifying work is missing, gather evidence of registration and ownership and contact the claims administrator. Strict deadlines apply; late claims normally need court approval.
https://secure.anthropiccopyrightsettlement.com/lookup
Step 5: Review rights splits and file or update your claim
Default splits for many trade books are 50/50 between authors and publishers unless the contract says otherwise or both parties elect a different percentage. Education works usually follow the publishing contract rather than the default.
Submit the claim form, select a payment method, and review the allocation carefully. Dispute incorrect percentages or missing rightsholders before the cutoff date. Undisputed allocations are paid as shown and generally cannot be reversed later. Keep copies of every confirmation.
Step 6: Block AI training crawlers with robots.txt
Reduce future unauthorized scraping of your web content by blocking known AI training crawlers while leaving ordinary search engines and AI search bots open if you want visibility. Edit the robots.txt file at the root of your domain.
# Block AI training crawlers
User-agent: GPTBot
Disallow: /
User-agent: ClaudeBot
Disallow: /
User-agent: Google-Extended
Disallow: /
User-agent: CCBot
Disallow: /
User-agent: Bytespider
Disallow: /
User-agent: Applebot-Extended
Disallow: /
# Allow everything else
User-agent: *
Allow: /
Sitemap: https://yourdomain.com/sitemap.xml
Verify the file is reachable at yourdomain.com/robots.txt. Robots.txt is a voluntary request that reputable companies honor; it is not a technical barrier against determined scrapers.
Step 7: Add clear copyright notices and set a monitoring routine
Place a visible copyright notice on every work and website page. Include the year of first publication and the name of the copyright owner. Embed copyright metadata in digital files where the format supports it.
Schedule a quarterly review of your inventory, new registrations, open settlement portals, and server logs. Subscribe to updates from the Authors Guild and claims administrators. Store all documents in a redundant, encrypted location so the next settlement is faster and more complete.
Common Mistakes
- Waiting until a settlement is announced to register—eligibility almost always requires timely registration that precedes or closely follows publication and the infringing use.
- Using the Single Application when the work has co-authors or is a work made for hire; the application will be rejected or delayed.
- Failing to disclose AI-generated material that is more than minimal; the registration can later be cancelled or weakened.
- Missing dispute deadlines in the allocation portal; incorrect undisputed splits are paid as listed.
- Blocking all AI-related user-agents in robots.txt and accidentally removing content from AI search results.
- Relying solely on robots.txt without also keeping strong registration and contract records.
Final Check
Confirm the following before you finish:
- Your works inventory is complete and includes any existing registration numbers and dates.
- Unregistered works have been filed through eCO with the correct application type and deposit copies.
- You have searched the current settlement Works List and recorded every match.
- Claims are submitted, allocations reviewed, and any disputes filed before deadlines.
- robots.txt is live and blocks the intended training crawlers.
- Copyright notices and supporting documents are organized and backed up.
Summary
Timely copyright registration through the eCO system, a complete works inventory, careful claim filing, and basic technical controls such as robots.txt give authors and publishers practical tools to collect their share of AI settlement funds and reduce future unauthorized training use. The same documentation process that lets rights holders receive full settlement proceeds works for anyone who keeps solid records.
Start today by searching the Copyright Public Records System and filing any missing registrations. Then check the official Works List for any open settlement, update your robots.txt, and set a quarterly reminder. These steps remain effective for every future AI copyright settlement and for protecting new works as they are published.

