News-July-06-2026

July 6, 2026 — Chinese consumer AI platforms began shutting down custom and humanlike agent features as national rules for anthropomorphic AI interaction services approached their July 15 effective date; NOTUS reported that a draft U.S. Treasury Department report warned federal financial regulators about systemic risks from the AI market; Illinois Governor JB Pritzker signed SB 315, creating a frontier-AI safety framework with independent third-party audits; and Reuters reported that CISA is using Anthropic's Mythos model to audit federal code repositories.

Contents

1. ByteDance and Alibaba disable custom AI agents ahead of China rules
2. Treasury analysts warn on AI market risks
3. Illinois enacts frontier-model safety audit law
4. CISA uses Anthropic Mythos for federal code audits

---

ByteDance and Alibaba disable custom AI agents ahead of China rules

ByteDance's Doubao and Alibaba's Qwen announced shutdowns of user-created or humanlike AI agent functions days before China's Interim Measures on anthropomorphic AI interaction services take effect on July 15. Doubao said its agent feature would go offline on July 15, while Qwen said humanlike interactive agents and user-created agents would be disabled on July 10 and broader agent functions would go offline by July 15. TechNode reported that both companies said users would retain temporary access to agent configurations and chat histories, but related data would no longer be recoverable after October 15, 2026. The Next Web reported that the rules target bots offering sustained emotional interaction while sparing workplace and productivity agents, and that Tencent had removed a similar Yuanbao feature in June.

The episode is significant for AI law because it shows Chinese platforms preemptively redesigning consumer-facing AI companion and role-play products to comply with an anthropomorphic-interaction regime before the rules formally take effect.

---

Treasury analysts warn on AI market risks

NOTUS reported that a draft Treasury Department report, prepared by career analysts for Treasury Secretary Scott Bessent, Federal Reserve Board Chair Kevin Warsh, and federal financial regulators, compares parts of the AI market to the dot-com bubble while warning that AI firms are more deeply embedded in the U.S. economy than dot-com predecessors. The report says a downturn in the AI market could affect stock markets, private credit, data-center financing, cloud providers, chip manufacturers, and utilities, even while noting that many leading AI companies are more mature and better capitalized than late-1990s dot-com firms. NOTUS framed the draft as a contrast with the administration's public pro-investment posture toward AI and reported that the document had been completed for weeks but was still awaiting formal approval before delivery to its intended audience and possible public release.

The development is significant for AI law because it suggests federal financial regulators are evaluating AI-related market concentration, infrastructure finance, and systemic-risk questions even as public AI policy remains focused on accelerating investment and deployment.

---

Illinois enacts frontier-model safety audit law

Illinois Governor JB Pritzker signed SB 315, the Artificial Intelligence Safety Measures Act, on July 6, 2026, establishing a state framework for large AI developers to identify, disclose, and mitigate catastrophic risks. The governor's office said the law requires covered developers to maintain public risk-management frameworks, provide transparency around frontier-model risks, create confidential reporting channels, protect employees who raise AI-safety concerns, and undergo regular independent third-party safety audits by qualified experts without financial conflicts of interest. CBS News Chicago reported that SB 315 targets the largest AI developers through thresholds including $500 million in revenue and a large computing measurement, and that OpenAI and Anthropic supported the bill during the legislative process. CBS News Chicago also reported that the bill passed the Illinois House 110-0 and the Illinois Senate 52-5 before Pritzker signed it.

The enactment is significant for AI law because Illinois moved from proposed frontier-model oversight to an enacted state safety regime centered on transparency, whistleblower reporting, critical-risk management, and third-party verification.

---

CISA uses Anthropic Mythos for federal code audits

Reuters reported that the Cybersecurity and Infrastructure Security Agency's Attack Surface Evaluation team is using Anthropic's Mythos model to audit government code repositories and has already found a large number of vulnerabilities. Reuters described the deployment as an example of a federal cyber agency using a frontier AI system for operational code review rather than merely studying AI policy in the abstract. The story is significant for AI law because it places AI procurement, federal cybersecurity operations, model reliability, and government oversight of AI-assisted vulnerability discovery into the same regulatory frame.