News-August-10-2026

August 10, 2026 — The Delhi High Court granted actor Tabu interim protection against unauthorized use of her identity through AI-generated content, deepfakes, and face morphing, and directed major online platforms to remove more than 150 identified URLs. In the United States, House Democrats asked Speaker Mike Johnson to bring major AI-company executives before Congress to testify under oath about recent model-driven cyber incidents. California also launched what the governor’s office described as a first-in-the-nation AI cyber-defense program to help protect critical infrastructure and state assets. The Ninth Circuit rejected interlocutory appeals by Meta and TikTok in consolidated youth-addiction litigation and denied Meta’s request to delay an August 12 trial brought by state attorneys general. South Korean lawmakers advanced a bill to detailed subcommittee review that would criminalize creating, distributing, or possessing AI-generated sexually explicit material depicting fictional people. Senator Bernie Sanders asked Meta, OpenAI, and Anthropic to pause AI development and warned that the Senate could regulate if the companies continued deploying increasingly capable systems at their current pace. Anthropic also outlined a worldwide machine-readable marking plan for supported Claude models and outputs as part of its implementation of the EU AI Act's Article 50 transparency code. South Australia also established a royal commission to examine AI’s effects on work, creative industries, education, public services, and other areas of society, with the inquiry expected to begin October 1 and report by July 1, 2027.

Contents

1. Delhi High Court orders takedown of AI and deepfake content using Tabu’s identity
2. House Democrats seek testimony from AI executives after cyber incidents
3. California launches AI cyber-defense program for critical infrastructure
4. Ninth Circuit lets youth-addiction suits proceed and denies Meta trial delay
5. Korea moves to criminalize AI-generated sexual content of fictional people
6. Sanders asks major AI labs to pause development
7. Anthropic sets worldwide Claude watermarking plan
8. South Australia establishes AI royal commission

Delhi High Court orders takedown of AI and deepfake content using Tabu’s identity

The Delhi High Court issued an ex parte interim injunction protecting Tabu’s name, image, voice, likeness, and other personality attributes from unauthorized exploitation, including through artificial intelligence, deepfakes, and face-morphing technology.

The order directed Google, Meta, X, Reddit, eBay, and other services to remove or disable access to more than 150 identified URLs; the principal platform-specific takedown directions were to be completed within 36 hours. The court also directed relevant platforms and service providers to disclose basic subscriber information and IP logs connected to specified accounts or content.

Justice Jyoti Singh registered the matter as a commercial suit, issued summons, and set December 7, 2026 as the return date for a discovery-related application; the interim injunction remains effective until the next hearing unless the court orders otherwise.

---

House Democrats seek testimony from AI executives after cyber incidents

A group of House Democrats asked House Speaker Mike Johnson to bring the leaders of OpenAI, Anthropic, and other major AI companies before Congress to testify under oath about recent cyber incidents involving company-developed models.

The August 10 letter, led by Representative Greg Casar of Texas, said Congress had not adequately responded to risks from AI development and also requested testimony from independent experts. The lawmakers described the recent incidents as serious and warned that they could foreshadow larger problems if increasingly capable models advance without regulation.

The signatories cannot compel testimony without action from Republican House leadership, so the letter is a request for a congressional hearing rather than a subpoena or scheduled appearance.

---

California launches AI cyber-defense program for critical infrastructure

Governor Gavin Newsom announced a first-in-the-nation AI cyber-defense program designed to help protect California’s critical infrastructure and state assets.

---

Ninth Circuit lets youth-addiction suits proceed and denies Meta trial delay

The U.S. Court of Appeals for the Ninth Circuit rejected interlocutory appeals by Meta and TikTok that sought to overturn rulings requiring the companies to face more than 3,000 consolidated federal lawsuits alleging that Meta, Google, TikTok, and Snap designed social-media products to addict young users.

The court held that Section 230 supplies a defense to liability rather than immunity from being sued, making the companies’ appeals premature before final judgment. The Ninth Circuit also denied Meta’s request to postpone an August 12 trial brought by 29 state attorneys general over alleged collection and use of children’s data, addictive platform design, and consumer-safety representations.

---

Korea moves to criminalize AI-generated sexual content of fictional people

A bill introduced by Representative Huh Young and nine other Democratic Party of Korea lawmakers advanced to detailed review by a subcommittee of the National Assembly’s Legislation and Judiciary Committee.

The proposal would amend South Korea’s sexual-crimes law to criminalize creating, distributing, or possessing AI-generated sexually explicit material depicting fictional people who could be mistaken for real individuals, with maximum penalties of seven years in prison or a 50 million won fine.

---

Sanders asks major AI labs to pause development

Senator Bernie Sanders sent a letter to the chief executives of Meta, OpenAI, and Anthropic asking them to pause AI development, saying recent model capabilities and cyber incidents showed that safety risks had reached a critical threshold.

Sanders warned that the Senate would pursue regulation if the companies continued deploying increasingly capable systems at their current pace; the letter was a policy demand, not a binding order or introduced bill.

---

Anthropic sets worldwide Claude watermarking plan

Anthropic said it signed the EU AI Act Article 50(2) Code of Practice on Transparency of AI-Generated Content and plans to add machine-readable marks to supported Claude-generated content.

Claude models launched in the European Union on or after August 2, 2026 will support marking from launch, while the company is working to extend support to earlier models during the statutory transition period; supported marks will apply wherever Claude is offered, including outside the EU.

Anthropic plans to embed imperceptible watermarks in generated text and use digitally signed C2PA provenance metadata for supported generated files. The company cautioned that detecting a mark would not conclusively establish that Claude authored the content, and the absence of a mark would not prove that AI was uninvolved.

See implementation tracker: EU AI Act Implementation Milestones 2026.

---

South Australia establishes AI royal commission

South Australian Premier Peter Malinauskas announced a royal commission into artificial intelligence to examine the technology’s effects on work, creative industries, education, public services, and other areas of society.

The inquiry is expected to begin on October 1, 2026, after terms of reference are settled, and to deliver its final report by July 1, 2027. The state plans to appoint three royal commissioners.

Malinauskas said the inquiry is intended to inform proactive government regulation while helping the state capture AI’s economic opportunities; he also warned that unchecked AI poses risks to society and the future of work.