Jump to content

AI Law Wiki News for June 8, 2026

From AI Law Wiki

AI-generated text. This page was generated using artificial intelligence.

June 8, 2026 — Today’s AI-law update covers the European Commission’s draft guidance on high-risk AI-system classification under the EU AI Act and Meta’s notice that an AI-assisted Instagram account-recovery vulnerability affected more than 20,000 accounts.[1][2][3]

Contents

  1. European Commission publishes draft guidelines on high-risk AI classification
  2. Meta discloses AI-assisted Instagram account-recovery breach
  3. Carreyrou defendants severed

European Commission publishes draft guidelines on high-risk AI classification

On May 19, 2026, the European Commission published draft, non-binding guidelines on classification of high-risk AI systems under the EU AI Act, with separate materials addressing general principles, Annex I regulated products, and Annex III high-risk use cases.[1][2] The draft guidance states that complex systems made up of multiple AI components, including agentic AI systems, should be assessed holistically when interacting components materially influence decisions in high-risk use cases.[1] For Annex I products, the guidance explains that Article 6(1) can apply whether an AI system is embedded in a product or provided separately, and that the AI Act’s own definition of a safety component controls the high-risk analysis.[1] For Annex III use cases, the guidance emphasizes intended purpose, takes a narrow view of the Article 6(3) filter, and discusses biometric systems, employment uses, creditworthiness systems, and fraud-detection exceptions.[1] The consultation on the draft guidance is open until June 23, 2026.[1][2]

Meta discloses AI-assisted Instagram account-recovery breach

Meta Platforms filed a data-breach notice with the Maine Attorney General stating that 20,225 people, including 30 Maine residents, were affected by unauthorized Instagram account access tied to a vulnerability in an AI-assisted account-recovery system.[3] The notice states that the incident date was April 17, 2026, that Meta discovered the issue on May 31, 2026, and that written notifications were sent on June 6, 2026.[3] Meta’s notice says that potentially involved data included contact information, dates of birth, photos, videos, stories, direct messages, account activity, profile information, and linked accounts, while also saying the company had no evidence that specific information was accessed.[3] Help Net Security reported that Meta disabled the AI-assisted recovery feature and required impacted users to reset passwords and re-authenticate through secure channels.[4]

Carreyrou defendants severed

Judge P. Casey Pitts severed Google, xAI, Perplexity, Apple and NVIDIA from Carreyrou v. Anthropic on June 8 (ECF 195). Dismissal from this action followed severance; it was not a decision approving their alleged copying.[5]

References

  1. ↑ 1.0 1.1 1.2 1.3 1.4 1.5 Inside Privacy, "EU AI Act Update: The European Commission Publishes Draft Guidelines on HRAIs"
  2. ↑ 2.0 2.1 2.2 European Commission, "Draft Commission Guidelines on the classification of high-risk AI systems"
  3. ↑ 3.0 3.1 3.2 3.3 Maine Attorney General data breach notification for Meta Platforms, Inc.
  4. ↑ Help Net Security, "Hackers used Meta’s AI support system to hijack over 20,000 Instagram accounts"
  5. ↑ N.D. Cal., Carreyrou v. Anthropic, No. 5:25-cv-10897-PCP, court docket entries reproduced by Justia; snapshot retrieved July 2, 2026
Retrieved from "https://ailawwiki.com/index.php?title=AI_Law_Wiki_News_for_June_8,_2026&oldid=3295"