News-June-15-2026

June 15, 2026 — U.S. officials and Anthropic prepared talks over a federal order restricting foreign-national access to Claude Fable 5 and Claude Mythos 5, reporting on Tesla Full Self-Driving submissions highlighted European regulatory scrutiny of AI-enabled vehicle-safety claims, a California federal judge denied Meta's motion to dismiss an adult-film copyright suit alleging BitTorrent downloads for AI training, and another California federal judge dismissed xAI's trade-secret case against OpenAI with prejudice.

Contents

1. Anthropic and White House prepare talks over Claude model access restrictions
2. Tesla and European regulators scrutinize Full Self-Driving safety data
3. Meta must face adult-film copyright suit over alleged AI-training downloads
4. xAI trade-secret case against OpenAI dismissed with prejudice

---

Anthropic and White House prepare talks over Claude model access restrictions

Business Today reported on June 15, 2026, that Anthropic staff were in Washington for a meeting with White House officials after the U.S. government ordered Anthropic to block access to Claude Fable 5 and Claude Mythos 5 for foreign nationals, whether inside or outside the United States. Silicon Republic reported that the order followed a late-Friday U.S. government directive requiring Anthropic to suspend access to the models by foreign nationals, including foreign-national Anthropic employees. Business Today described Claude Fable 5 as a model that had been publicly available for heavy-duty knowledge work and described Mythos 5 as a model whose access had been restricted to a limited group of organizations. Silicon Republic characterized the order as an export-control directive and reported that senior Anthropic technical staff were expected to meet White House officials in Washington to resolve the dispute.

TechCrunch reported that dozens of cybersecurity experts urged the White House to remove the restrictions, arguing that limiting Fable and Mythos access would reduce defenders’ ability to secure software and products. The open letter framed the restriction as a cyber-defense risk and asked the administration to adopt transparent AI cyber-protection rules rather than broad access limits.

The development is significant for AI law because it applies federal national-security and export-control tools directly to access to frontier AI systems, including access by employees based on nationality. It also places model-access governance, employee access controls, and cross-border AI deployment within the same regulatory frame as export restrictions on sensitive technologies.

---

Tesla and European regulators scrutinize Full Self-Driving safety data

The Next Web reported on June 15, 2026, that Reuters had reviewed documents showing Tesla presented European regulators with Full Self-Driving safety data that traffic-safety researchers described as misleading. The report said Tesla compared driver-supervised Full Self-Driving crash rates with overall U.S. crash statistics rather than with a like-for-like baseline for similar driving conditions. The Next Web reported that the reviewed materials concerned submissions to European regulators as Tesla sought approvals connected to Full Self-Driving deployment in Europe.

The development is significant for AI law because it connects automated-driving safety claims, regulator-facing evidence, and consumer-protection concerns about how AI-enabled vehicle capabilities are represented to oversight authorities.

---

Meta must face adult-film copyright suit over alleged AI-training downloads

TorrentFreak reported on June 15, 2026, that a California federal judge denied Meta's motion to dismiss a copyright suit by adult-content producers Strike 3 Holdings and Counterlife Media over alleged BitTorrent downloads of adult films. The report said the complaint accused Meta of using adult films to assist AI model training, but the court held that the plaintiffs did not need to prove AI-training use at the motion-to-dismiss stage because the alleged torrenting itself stated an infringement theory. TorrentFreak reported that Meta had argued any downloads were for personal use rather than AI training, and that the court did not accept dismissal on that basis at this stage. The court order made available by TorrentFreak denied Meta's motion to dismiss and allowed the copyright claims to proceed.

The development is significant for AI law because it permits a copyright-infringement case tied to alleged AI-training data acquisition to proceed even without requiring the plaintiffs to prove the downstream training use in their pleadings.

---

xAI trade-secret case against OpenAI dismissed with prejudice

The Next Web reported on June 15, 2026, that U.S. District Judge Rita Lin dismissed xAI's trade-secret lawsuit against OpenAI with prejudice, meaning xAI cannot bring the same claims again. The court's order stated that xAI alleged former senior engineer Xuechen Li took confidential materials before joining OpenAI, but held that the amended complaint still did not plausibly allege that OpenAI induced misappropriation or knew trade secrets were being disclosed during hiring. The order said xAI's allegations about OpenAI's desire to compete in the chatbot market and its hiring of former xAI employees did not bridge the gap between employee mobility and OpenAI's own trade-secret liability. The Next Web reported that the dismissal was Musk-linked xAI's second OpenAI-related court loss in four weeks after a separate jury verdict rejected claims over OpenAI's charitable mission.

The development is significant for AI law because it limits trade-secret claims based on hiring from competing frontier-model companies unless the plaintiff can plead facts tying the new employer to inducement, knowledge, or use of protected information. It also underscores how AI-sector employee mobility disputes may turn on ordinary trade-secret pleading standards rather than on the competitive stakes of model development alone.