Recent key developments in AI Law in the United States

AI is the hot ticket item of 2026 and not only the technology landscape around AI is changing rapidly, so is the legal landscape. This newsletter is not intended to provide a comprehensive overview of all the recent legal developments concerning AI in the US but is an intentional selection of state laws deemed to be of particular interest to European companies doing business in the USA, either directly or through US subsidiaries.

While the federal government under President Trump initially adopted a laissez-faire approach to AI, after public pushback in the beginning of 2026, the administration has increased regulatory oversight through federal cybersecurity reviews by mandating the creation of an AI cybersecurity clearing house by the end of June/beginning of July.

On June 2nd, 2026, President Trump issued an executive order with the title “Prompting Advanced Artificial Intelligence Innovation and Security” which direct agencies to develop a classified benchmarking process that determine the cyber capabilities of AI models as a “covered frontier model” and to design a framework for the federal government to access covered frontier models for up to 30 days before broader release to the public. “While the government evaluation as a covered frontier model is voluntary at the moment, the voluntary framework could facilitate a formal AI oversight process down the road.

State Law Trends

States are aggressively adopting AI laws resulting in a fragmented patchwork of AI laws. Key legislative trends on the state level include rules for AI chatbots and protection against deepfakes, “frontier” model safety frameworks and transparency mandates for automated employment decisions.

Digital Replicas and Deepfakes

On the state level, there are several states that have enacted laws governing the US of AI Replicas and Synthetic Performers, most notably New York and California.

NY Synthetic Performer Law

Since June 9, 2026, the NY Synthetic Performer Law requires anyone who produces or creates an advertisement to identify if the ad includes AI-generated synthetic performers.

An AI-generated synthetic performer is defined as a digitally created asset made using generative AI or software algorithms that creates the impression of a human performer but is not recognizable as any identifiable natural person.
The law does not contain any information on the exact font, size, or length for disclaimers but states that the disclaimer must be “conspicuous”. The term “conspicuous” is an established legal term in the context of warranty disclaimers.
Since there is no apparent guidance in the statute nor case law on this question yet, it can be inferred from the term “conspicuous” that the disclaimer must be noticeable, legible, and understandable to a reasonable consumer at the point of viewing.

• Size: Proportional to the size of the advertisement. It should be large enough to be easily readable on a small mobile device.
• Placement: Placed directly on or near the synthetic performer rather than buried in fine print at the bottom of the ad.
• Duration: Present on-screen for the duration of the visual performance, not just flashing for a second.
• Audio: Paired with an audio disclosure (“The people in this ad are AI-generated”) if there is a voiceover.
• Clarity: Written in clear, plain language (e.g., “This ad features synthetic performers” or “AI-generated performers”) rather than confusing legal jargon.

Violations are subject to penalties of $1,000 for the first violation and $5,000 for subsequent violations.

California’s Digital Replica Act

California’s Digital Replica Act refers to two landmark artificial intelligence laws that protects living and deceased performers from having AI and digital replicas replace their actual work. A digital replica is defined as “a computer-generated, highly realistic electronic representation that is readily identifiable as the voice or visual likeness of an individual that is embodied” in a digital work (Sound recording, image, audiovisual work) “in which the individual did not actually perform or in which they did perform but the fundamental character of the performance has been materially altered”. Signed into law and effective for performances after January 1, 2025, the legislation voids any contract provision allowing an AI replica to substitute for a real worker, unless the worker had proper legal or union representation.

Frontier Models & Transparency

Another trend on the state level is to regulate large frontier AI models due to their scope, impact and potential risks. These efforts are led by New York and California and require developers to implement safety protocols, periodically publish safety protocols and report dangerous incidents.

The California AI Transparency Act

The California AI Transparency Act, which goes into effect on August 2, 2026, and regulates large generative AI providers (with over 1 Million users) to protect consumers from deepfakes and misinformation by introducing mandatory watermarking, AI detection tools, and provenance labeling as well as civil penalties for non-compliance of up to one million US dollars for non-compliance. Given the high number of large AI developers based in California, this new law can be expected to have global reach.

Chatbots and Generative AI contents

A common concern in the state laws governing AI is Generative AI transparency and consumer protection, many states enacted disclosure obligation for chatbots to address the fear that users might not be aware that they are communicating with a chatbot instead of a human being. California and New York’s laws are already in effect, whereas the laws in Colorado, Idaho, Nebraska, Oregon and Washington will go in effect later this year or in some states in July 2027.

Automated Decision-Making and Employment

There are several states that have already enacted laws regulating automated decision-making in employment matters, most notably Colorado, Illinois, California and Texas. The most comprehensive state law is the “Colorado AI Act” which requires employers to notify employees when using AI for hiring or promotions and to conduct bias assessments as well as to provide opportunities to correct data.

Outlook

The outlook for AI regulation in the US reflects a significant shift towards AI oversight and – for now – a fragmented AI legal framework. The Trump administration is pursuing a “minimally burdensome” national AI regulatory framework through executive action and legislative recommendations released in March 2026. However, states continue to enact their own AI laws, creating ongoing tension between federal preemption efforts and state regulatory activity. No comprehensive federal AI law currently exists, though multiple federal agencies are actively working to establish unified standards across key areas such as employment, consumer protection, and critical infrastructure.

Unless and until Congress enacts a national AI standard or framework that preempts state law, employers and other businesses must continue to comply with a patchwork of evolving state and local laws governing the use and deployment of AI tools. Organizations implementing AI must be aware of laws in all jurisdictions where they operate, how they differ, and what responsibilities they impose.

We will continue to monitor the regulatory AI framework in the US. For more information, please contact us.

Dr. Annamaria Hachmeister