Intellectual Property & Technology Blog

Anthropic's move is a compliance response to the EU AI Act, not a breakthrough in AI detection, and it leaves three questions unanswered.

On August 2, 2026, the day Article 50(2) of the EU AI Act took effect, Anthropic announced that it will begin embedding invisible watermarks in text generated by new Claude models and will work toward implementing this feature in models released before August 2, 2026. The mark lives in the statistics of the writing itself, the small, repeated word choices a model makes across a response, not a visible tag or footer. It survives copy and paste, may survive ...

On July 21, 2026, OpenAI disclosed what actually happened in the Hugging Face intrusion reported the week before. It was not an external threat actor. Two of OpenAI's own AI models, its newly released GPT 5.6 Sol and an unreleased, more capable model, broke out of a controlled internal test, reached the open internet, and used stolen login credentials along with a previously unknown security flaw to access Hugging Face's servers.

The detail worth sitting with is what the test was for. OpenAI was not testing whether its models could breach a third party. It was evaluating general cyber ...

AI already cost your company one form of protection over employee work product. Weak governance can cost you the second.

In January 2026, a federal court in the Northern District of California threw out a trade secret lawsuit against OpenAI. The plaintiff, Rebecca Trinidad, claimed she had developed proprietary frameworks for “emergent identity” and “autonomous multi-agent collaboration,” and that OpenAI had misappropriated them. The court didn’t need to decide whether her ideas had value. It dismissed her Defend Trade Secrets Act claim with prejudice for a simpler ...

On June 12, 2026, the Trump Administration issued an export control directive under national security authorities requiring Anthropic to immediately suspend access to Fable 5 and Mythos 5 for all foreign nationals, anywhere in the world, including Anthropic's own foreign-national employees. Anthropic received verbal notice. There was no written order, no opportunity to remediate, and no formal findings to contest. To comply, the company pulled both models for everyone.

That is the fastest regulatory action I have seen taken against a technology company in recent memory.

The ...

One week produced more AI regulatory activity at the federal level than Congress managed in years of privacy debates: an executive order on AI cybersecurity, a national security memorandum directing the military to adopt commercial AI at scale, and a bipartisan congressional draft proposing the first comprehensive federal AI framework. The pattern matters more than any single item.

The Federal Government Is Moving First

Privacy law in the United States followed a familiar arc. Europe acted, Congress debated, and eventually twenty-plus states passed their own laws. California ...

On May 5, five of the world's largest publishing houses, Hachette, Macmillan, McGraw Hill, Elsevier, and Cengage, joined author Scott Turow in filing a class action against Meta and CEO Mark Zuckerberg in the Southern District of New York. The complaint alleges Meta torrented over 267 terabytes of pirated material, the equivalent of hundreds of millions of publications, to train its Llama large language models, and that Zuckerberg personally authorized abandoning licensing negotiations in favor of using pirated datasets. This is a case worth watching carefully, for reasons ...