Kilpatrick’s Betsy Bengtson, Darin Brown, and Andrea LaFrance recently presented “A Trademark Practitioner’s Guide to Using AI: Guidelines, Use Cases, and Ethical Considerations” during the firm’s annual “SKI-LE” in Breckenridge, Colorado. Betsy, Darin, and Andrea discussed how artificial intelligence (AI) is rapidly transforming the practice of law and offered attorneys powerful tools to increase efficiency, accuracy, and client value. However, they pointed out that integrating AI into legal workflows requires a nuanced understanding of ethical obligations, internal protocols, and potential risks. The session provided trademark practitioners with comprehensive and practical guidance on responsibly utilizing AI throughout the trademark lifecycle.
Key takeaways from the presentation include:
1. Fair Use in the AI Trenches: Market Harm Matters for AI Inputs: Courts are scrutinizing whether AI model training on copyrighted or trademarked content constitutes fair use, with a growing emphasis on whether the use is transformative and whether rightsholders can show real market harm. Recent cases demonstrate that use of pirated materials or direct commercial substitution weighs heavily against fair use, making the source of training data and evidence of market impact critical to litigation outcomes.
2. AI Outputs: New Frontiers for Infringement Liability: AI-generated outputs can expose developers and platforms to direct and contributory infringement risk. Courts have allowed claims to proceed where end users create infringing content, and the platform knowingly enables or materially contributes to that conduct.
3. Generative AI: Dilution, Confusion, and Trademark Turbulence: Generative AI can inadvertently produce content that confuses consumers, dilutes famous marks, or creates misleading attributions. Courts have allowed dilution, false endorsement, and trade dress claims to proceed, and are considering whether AI-generated articles or images can themselves serve as evidence of consumer confusion.
4. EU AI Act: A Compliance Wake-Up Call: The EU AI Act sets a global benchmark for AI regulation, mandating disclosure of synthetic content, compliance with IP laws, banning “unacceptable risk” systems (like social scoring or exploitative emotion recognition), and strict penalties for violations. So even if something is technologically feasible or patentable, it may not be deployable in the EU.
5. U.S. AI Law: The Regulatory Ground Shifts: U.S. law is evolving quickly, with new statutes like the TAKE IT DOWN Act targeting deepfakes and requiring content removal within 48 hours upon notice, and bills under consideration poised to create federal rights in voice and image, increase transparency in AI training, and limit AI procurement from foreign adversaries, signaling a rapidly tightening regulatory environment.
6. AI and Privilege: Confidentiality Not Guaranteed: Clients should be aware that AI-generated documents are not automatically protected by attorney-client privilege. See United States v. Heppner, No. 25-cr-00503-JSR (S.D.N.Y. Feb. 10, 2026).
