On September 1, 2026, the U.S. Department of Justice filed a position before a federal court in New York in favor of OpenAI and Microsoft in their lawsuit against the New York Times. The heart of the debate is fair use, a concept in American law that may allow certain transformative uses of protected works without prior licensing.
The professional stakes go beyond the press: companies, law firms, universities, and administrations are already using artificial intelligence (AI) to produce notes, contracts, summaries, or analyses. In Libreville, a corporate lawyer who asks a tool to summarize OHADA (Organization for the Harmonization of Business Law in Africa) contracts must therefore verify what is exploitable, confidential, and legally safe.
Recent articles
Google brings YouTube closer to messaging: opportunity and risk
How to use artificial intelligence in finance?
AI at work: how to succeed without disorganizing?
Certificate
Certificat Professionnel en Droit de l’Energie
What is it, concretely?
To create a generative AI like ChatGPT, a company trains its systems on very large volumes of texts. Some may be protected by copyright: articles, books, archives, commented decisions, standard contracts, or professional documents. Before reusing protected content, an organization often had to obtain a license or identify a legal exception. Here, the U.S. government tells the judge that training an AI can be transformative enough not to be automatically illegal.
The key point: this position is neither a law, nor a regulation, nor a definitive judicial victory for OpenAI. Judicial and media sources describe a procedural act in a private litigation. The judge will still have to assess the facts. Above all, this does not resolve other risks: data provenance, possible reproduction of protected passages, confidentiality of documents sent to the tool, personal data protection, and applicable law outside the United States.
Concrete case: what to do and what not to do
Questions to Ask Before Acting
- If our organization uses an AI trained or fed by protected content, who bears the legal risk: the provider, us, or both?
- Does what this case changes in our profession represent a mere technical possibility, or an obligation for documentary and contractual governance?
- Do the documents sent to the tool contain personal data, business secrets, HR information, or elements covered by professional secrecy?
- Have we understood the provider's clauses on data retention, reuse, deletion, and location?
- Are the produced responses verified by a lawyer, tax advisor, HR manager, or competent professional before external use?
- Do our contracts with clients allow the use of AI tools and specify confidentiality limits?
- In case of a dispute, can we explain how the note, contract, or decision was produced, reviewed, and validated?
UNIVGA Viewpoint
Sources
- Associated Press, Trump administration backs OpenAI in New York Times copyright case
- Agency, DOJ Tells Court AI Training on Copyrighted Text Is Fair Use
- Justia, The New York Times Company v. Microsoft Corporation et al.
- U.S. Copyright Office, Copyright and Artificial Intelligence, Part 3
- DataReportal, Digital 2026: Gabon
- APDPVP, Authority for the Protection of Personal Data and Privacy
French 
English 