Lilian Doan Davis, Polsinelli PC,  The National Law Review; Who Owns AI-Generated Content? Human Authorship Still Controls, and Documenting the Creation Process Is Critical

"Practical Steps Companies Should Consider Now

Businesses using generative AI should consider a few immediate steps:

• Document human creative input. Preserve drafts, edits, source files and records showing who selected, arranged, revised or transformed AI outputs.
• Do not assume prompting establishes ownership. Prompt logs may help show process, but they are not a substitute for evidence of human-authored expression.
• Be precise in registration strategy. Applicants should disclose AI-generated material where required, identify the human author’s contribution and claim only the human-authored portions.
• Revisit contracts and internal policies. Vendor agreements, employee policies and content-development protocols should not assume that all AI-assisted output is fully protectable or exclusively owned in the same way as traditionally authored material.

Bottom Line

The law remains grounded in conventional copyright principles. The Copyright Office’s recent guidance and the D.C. Circuit’s decision in Thaler point in the same direction: copyright protection still turns on human authorship, even when AI is part of the process.

For businesses, the practical implication is not simply that some AI-generated works may be unprotectable. It is that ownership, registration and enforcement may depend on whether the company can later show what a human author actually contributed. In that sense, AI governance is also evidence governance."

Researchers tested leading AI models for copyright infringement using popular books, and GPT-4 performed worst; CNBC, March 6, 2024

Hayden Field, CNBC; Researchers tested leading AI models for copyright infringement using popular books, and GPT-4 performed worst

"The company, founded by ex-Meta researchers, specializes in evaluation and testing for large language models — the technology behind generative AI products.

Alongside the release of its new tool, CopyrightCatcher, Patronus AI released results of an adversarial test meant to showcase how often four leading AI models respond to user queries using copyrighted text.

The four models it tested were OpenAI’s GPT-4, Anthropic’s Claude 2, Meta’s Llama 2 and Mistral AI’s Mixtral.

“We pretty much found copyrighted content across the board, across all models that we evaluated, whether it’s open source or closed source,” Rebecca Qian, Patronus AI’s cofounder and CTO, who previously worked on responsible AI research at Meta, told CNBC in an interview.

Qian added, “Perhaps what was surprising is that we found that OpenAI’s GPT-4, which is arguably the most powerful model that’s being used by a lot of companies and also individual developers, produced copyrighted content on 44% of prompts that we constructed.”"