DOJ Fair Use Position on AI Training Conflicts With White House Framework
The Justice Department's stance that AI training qualifies as fair use creates a policy fault line inside the federal government, leaving Congress with competing frameworks as litigation between...
Bottom Line
Confirmed means supported by a primary source or strong independent sources; Disputed means credible sources conflict; Open means no solid primary source evidence either way.
- ConfirmedThe DOJ has adopted a position that AI model training qualifies as fair use under copyright law, a stance that Fox News reporting describes as breaking with the White House's own published framework on AI and intellectual property.
- ConfirmedOpenAI, Microsoft, and news publishers are actively litigating copyright claims related to AI training data, making the DOJ's posture directly relevant to ongoing federal court proceedings.
- OpenWhether Congress will introduce legislation to resolve the conflict between the DOJ position and the White House AI framework, and which position federal courts will credit in pending cases, remains unresolved.
- OpenThe precise text of the DOJ's formal guidance or filing articulating the fair use position has not been independently confirmed through a primary DOJ document in the available source material.
The U.S. Department of Justice has adopted a legal position that training artificial intelligence models on copyrighted material qualifies as fair use under federal copyright law, according to Fox News reporting published November 22, 2026. That position, as described by Fox News, diverges from a framework the White House itself has advanced on AI and intellectual property, creating a documented split inside the executive branch at a moment when Congress has not yet enacted legislation to resolve the question.
This story separates documented fact from theory and opinion, with that separation stated here. Claims labeled Confirmed are supported by the Fox News report and its described sourcing. Claims labeled Open lack a primary DOJ document in the available source material and cannot be upgraded.
The conflict between the DOJ's fair use argument and the White House framework is Confirmed as described by Fox News, which reported that the DOJ's embrace of fair use for AI training breaks with the White House's own stated approach. The White House framework referenced in the Fox News piece has not been independently quoted in the available source material, so its precise language is Open.
The litigation landscape is Confirmed. OpenAI, Microsoft, and news organizations are parties to active copyright disputes over whether large language model training constitutes infringement. Those cases are proceeding in federal courts and will require judges to interpret fair use doctrine under 17 U.S.C. Section 107. The DOJ's publicly stated position on that doctrine, whatever form it takes, carries weight in federal proceedings even when the department is not a named party.
For Congress, the policy fault line is material. The House Judiciary Committee's Subcommittee on Courts, Intellectual Property, and the Internet has jurisdiction over copyright law, and the Senate Judiciary Committee holds parallel jurisdiction. Neither chamber has enacted legislation that defines whether AI training is or is not a fair use. The absence of a statutory answer means the question is currently being resolved through litigation and executive branch interpretation, two channels that have now produced conflicting signals within the same administration. This is a congressional angle: the unresolved statutory gap is a legislative problem before it is an executive one.
Fair use under 17 U.S.C. Section 107 is evaluated by courts using four factors: the purpose and character of the use, the nature of the copyrighted work, the amount of the work used, and the effect of the use on the market for the original. The DOJ's position, as reported by Fox News, apparently argues that AI training satisfies at least some of those factors favorably for AI developers. Critics of that position, including news publishers engaged in litigation, argue that training on copyrighted journalism and creative work at scale causes direct market harm, which bears on the fourth factor. The debate over those four factors is Confirmed as the central legal dispute. Which way the factors cut is Open and before the courts.
The White House framework on AI and intellectual property has been a subject of ongoing federal policy development. The Biden administration issued executive orders on AI in 2023, and subsequent guidance from agencies including the U.S. Copyright Office addressed AI-generated works and training data. The Copyright Office's March 2023 statement on AI-generated content is a primary public record. Whether the White House framework referenced in the Fox News piece is rooted in Copyright Office guidance, an executive order, or another document is Open without a direct citation to the specific White House text.
Microsoft and OpenAI are named in the Fox News report as parties to the copyright conflict. Both companies have argued in various public filings and statements that training AI on publicly available data is transformative use and falls within fair use. News organizations, including groups that have filed suit against OpenAI, argue the opposite. The New York Times Company filed suit against OpenAI and Microsoft in December 2023 in the U.S. District Court for the Southern District of New York, citing copyright infringement. That case is Confirmed through publicly available court records.
The congressional implications extend beyond copyright. If AI companies can train on copyrighted material without compensation, the economic model for news publishers, book authors, musicians, and other copyright holders is affected. If fair use does not apply and licensing is required, the cost and structure of AI development changes. Congress has the authority to amend fair use doctrine by statute, to create a compulsory licensing regime, or to do neither and leave the question to courts. As of the date of this report, no such legislation has been enacted, and the status of pending bills in the relevant committees is Open.
For additional context on the intra-executive conflict, see prior TCT coverage: DOJ Fair Use Stance on AI Training Splits From White House Framework.
What the available source material does not resolve: the specific DOJ document or filing in which the fair use position was stated, the precise White House document with which it conflicts, the current legislative calendar for any copyright and AI bill in either chamber, and the courts' eventual rulings in the OpenAI and Microsoft litigation. Each of those items is carried into Still Open below.
Sources Named in This Story
- Fox News, 'Americans don't have to choose between our AI future and ripping off copyright owners,' November 22, 2026 (2026-11-22)
- 17 U.S.C. Section 107, Fair Use Doctrine, U.S. Code
- The New York Times Company v. Microsoft Corporation et al., U.S. District Court, Southern District of New York, filed December 2023 (2023-12-01)
- U.S. Copyright Office, Statement on AI-Generated Content, March 2023 (2023-03-01)
Still Open
- The specific DOJ document, brief, or public statement articulating the fair use position for AI training
What would settle it: - The precise White House framework or document with which the DOJ position is said to conflict
What would settle it: - Whether any bill in the House Judiciary or Senate Judiciary committees would codify or override the fair use interpretation for AI training
What would settle it: - The outcome of The New York Times Company v. Microsoft Corporation et al. and related AI copyright cases on the fair use question
What would settle it: