Generative AI raises three separate IP questions that are often blurred together: whether training on protected material is lawful, whether output infringes someone else's rights, and whether anyone owns the output at all. Courts have begun answering pieces of each, often in narrow and fact-specific ways. This course teaches you to separate the questions, read the leading decisions accurately, and turn them into practical rules for your organization.
You will cover the human authorship requirement (Thaler v. Perlmutter and the US Copyright Office's Copyright and Artificial Intelligence reports, including the pre-publication Part 3 on training), registration with AI-generated material, the fair use analysis in Bartz v. Anthropic, Kadrey v. Meta and Thomson Reuters v. Ross (including the Third Circuit's September 2026 affirmance), the significance of how training data was acquired, the EU text and data mining opt-out, Getty Images v. Stability AI in the Northern District of California (where most of Getty's copyright and trademark claims survived a motion to dismiss in April 2026) and the territorial question of models trained abroad, and GEMA v. OpenAI in Germany. You will then assess output risk from memorization, characters, trademarks and likeness, open-source contamination in generated code, and compare model licenses (Apache 2.0, which Google adopted for Gemma 4 in April 2026, the Llama community license, the custom terms on earlier Gemma releases, and OpenRAIL) and provider indemnity programs. The course ends with EU AI Act GPAI copyright duties and a clearance workflow. Case status is stated as of October 2026 and many matters remain on appeal. It is educational, not legal advice; the Wirewalk Institute certificate attests to completion and assessment of this course only.