Copyright Registration Accomplished!
This third blog in the series describes the final stage in the process in securing copyright registration for a PowerPoint presentation titled “Copyright in the Age of Artificial Intelligence.” The PowerPoint slides contained human-authored content and both AI-Generated and AI-Assisted content. Copyright registration requires disclosure of AI-generated content.
The differences in the copyrightability of human-authored v. AI-created content is the subject of Part 4 in this series.
The US Copyright Office (USCO) policy indicates that “applicants have a duty to disclose the inclusion of AI-generated content in a work submitted for registration and to provide a brief explanation of the human author’s contributions to the work.” This policy is identified in “Copyright Registration Guidance: Works Containing Material Generated by Artificial Intelligence” published on March 16, 2023.
This publication was released five months after the USCO cancelled the copyright registration given to Kris Kashtanova for her graphic novel titled “Zarya of the Dawn” because she failed to adequately disclosure the use of artificial intelligence in her application. She had used Midjourney AI to create the graphics for the book.
Kashtanova then submitted additional information to the USCO examiner who was processing her application. The USCO sent Kashtanova a letter in February 2023 that she was granted copyright protection for “the Work’s text as well as the selection, coordination, and arrangement of the Work’s written and visual elements.”
Human-authored v. AI-created content
We submitted an eCO (Electronic Copyright Office) standard application was submitted in mid-February to register the PowerPoint presentation. We also submitted a Change Request to improve the disclosure of AI-generated content in the work. A reply to the Change Request was quickly received from the USCO. It referred us to the document cited above.
This reply provided key clarification that information provided on the Limitation of Claims page on the application should only identify the types of expression such as “text, “two-dimensional artwork,” and “photographs.” The reply also provided detailed instructions for how and what to include in the change request, which we submitted with the new language.
Within hours an automated reply was received indicating our request had been brought to the attention of the registration specialist who was processing our application. The specialist followed up with emails seeking clarification about how we used AI. He was attempting to clearly identify what was human-authored content and what was both AI-Generated and AI-Assisted content in the submission. The vocabulary to accurately described how AI was used was still unclear to us.
The specialist requested a phone call and a Zoom session that gave us the opportunity to verbalize AI uses such as fact-checking and researching topics. This use of AI should have been referred to as AI-Assisted content, which might be copyrightable under certain circumstances. The slides also contain images generated by AI platforms only using prompts, which should have been referred to as AI-Generated content.
Publication of the Presentation
The remaining issue with the application was its status as a published or unpublished work. “Publication” has a technical meaning in US copyright law. The two times that the PowerPoint presentation was made and recorded prior to this point were not “published” because the recordings were only available to members of the associations that hosted the presentations. The USCO says “Generally, publication occurs on the date on which copies of the work are first made available to the public.”
We delayed finalization of the application AI-because another presentation was going to be recorded and made public on YouTube in only a few days. We submitted a copy of the recording of the PowerPoint presentation, including the audio from the presentation in lieu of the original slide set. The publication date was established when the recording was made public.
And the copyright registration was successfully completed in a matter of hours. We were confident that the audio portion of the presentation would be granted copyright protection as “text,” but we did not know if elements of the slides would be granted copyright protection. It was a question of human-authored v. AI-created content.
This is the exact language the “Certificate of Registration” that we received:
Pre-existing Material: preexisting text, artwork, and photographs; text and artwork generated by artificial intelligence.
Basis of Claim: text, photograph(s), compilation of artwork generated by artificial intelligence.
The reference to “text” included the audio recording and text on the slides that may reflect AI-Assistance such as research and fact-checking.
In conclusion, the process of applying for copyright registration greatly informed our understanding of the copyrightability of AI-generated and AI-assisted materials. That is the topic of the next blog in this series.
@TheCopyrightDetective is the handle used to identify Dr. C. Daniel Miller on social media, and the name is federally registered as a trademark.
#TheCopyrightDetective