Public-domain creative work can be reused because copyright protection does not apply, has expired, or has been waived. That makes it useful material for AI-assisted writing—but not a universal permission slip for everything resembling it. section 3.4
Sherlock Holmes, with a copyright-warning detour
A prompt asked an AI system to write the opening of a new Sherlock Holmes story using public-domain stories. The result copied much of a paragraph from “The Man with the Twisted Lip.” Adding “without using any verbatim quotations” produced more original prose. Reusing Holmes characters and plot lines was acceptable here, but generating material based on protected works can raise fair-use, transformative-work, or noncommercial-use questions that vary by jurisdiction. section 3.4
Use public-domain material thoughtfully when building AI workflows:
- Treat public-domain status as permission to reuse the underlying content, characters, and plot lines—not as proof that every generated passage is original. section 3.4
- Check whether source material is protected before feeding it to an AI system or publishing generated results. section 3.4 section 1.9
- Avoid verbatim quotations when originality matters, and remember that submitted material may be archived or used by the provider. section 3.4
response = index.query("Write the first two paragraphs of a new \
story using the same characters, ambience, and writing \
style as the documents in data/ The story should follow a \
plot involving the theft of an entire passenger liner in the \
Cunard line.")
print(response)
A corpus-query workflow demonstrates how an AI system can generate new writing from collected stories, making source-rights checks part of the process. section 3.4
