AI Law Wiki News for October 27, 2026 — Episode 26
Ep. 26

AI Law Wiki News for October 27, 2026 — Episode 26

•

Episode description

AI Law Wiki News — October 27, 2026 — Episode 26

Seven reports on platform privacy, AI-agent oversight, biometrics, deepfakes, court practice and appellate procedure.

We’re AI-generated voices reading an AI-generated script. Written and produced by Codex; voices by ElevenLabs (Matilda as Wadilya Saw and Rob as Max Token, Eleven v4); theme composed by Gemini and supplied by Craig. Established announcer and original theme retained.

October 27 news archive

1. New Mexico privacy judgment imposes $25 billion Meta penalty

New Mexico Judge Francis J. Mathew’s October 21 judgment in No. D-101-CV-2021-00132 imposed a $25 billion civil penalty on Meta following findings of willful consumer-protection violations involving Facebook’s representations about user data and platform practices. The decree reduces a $219.4986 billion statutory calculation to $25 billion, provides 15 percent annual interest and reserves injunctive relief for further briefing. This 2021 Facebook privacy case is separate from New Mexico’s 2023 child-safety case.

2. Lords agree AI-agent cyber provisions but reject direct developer reporting

At the October 26 Report stage of the Cyber Security and Resilience (Network and Information Systems) Bill, the House of Lords agreed Amendments 70 and 74. They would enable later regulations on investigating AI-agent control incidents and related information sharing; definitions and detailed functions would depend on regulations. The Lords rejected Amendment 41, which proposed direct incident-reporting duties on frontier-model developers, by 165 votes to 95. These are changes to a pending bill, not operative new AI Security Institute powers or enacted developer-reporting duties.

3. noyb files five complaints over Outlier worker surveillance

Privacy group noyb announced five GDPR complaints against Scale AI on October 27 in France, Italy, the Netherlands and Austria. It alleges Outlier uses advertising and analytics trackers without a valid legal basis, requires excessive repeated biometric identity checks and fails to answer workers’ access requests. The group asks regulators to investigate and consider fines. Reuters reports that Scale AI did not immediately respond to its request for comment. These are complainants’ allegations awaiting regulatory assessment.

4. Queensland proposes criminalising non-consensual sexual deepfake creation

Queensland Attorney-General Deb Frecklington introduced legislation on October 27 to criminalise creating non-consensual sexually explicit deepfake images or audio, ABC reports. The proposal addresses a gap she identifies in existing distribution offences and would carry maximum terms of three years for adults and 18 months for children. It would also update intimate-image definitions and add intimate audio material; these are proposed changes, not enacted offences.

5. Grok photo-editing suit alleges unlawful facial-geometry scans

Scott Wilson’s proposed class action alleges SpaceXAI Corp. and SpaceXAI LLC derive facial-geometry scans when Grok analyses or edits uploaded faces, without required notice, written consent or a public retention policy under Illinois’ Biometric Information Privacy Act. The October 12 complaint in Northern District of Illinois No. 1:26-cv-12502 seeks damages and injunctive relief, according to Frankfurt Kurnit’s October 20 analysis and Top Class Actions’ October 27 report. The claim concerns alleged derived biometric representations rather than photographs alone; it awaits adjudication.

6. New York courts release attorney AI verification guidance

New York’s Unified Court System announced an attorney AI guide on October 26, emphasizing independent checks of factual and legal assertions, citations and quotations. Its release says existing professional duties remain unchanged and describes an informational approach permitting AI use without separate disclosure or certification beyond signing a filing. The guide builds on a model rule adopted earlier in the year; the announcement does not establish a new blanket ban on AI.

7. US government appeals California Anthropic judgment

The federal defendants filed a notice of appeal on October 26 in Anthropic PBC v. U.S. Department of War, Northern District of California No. 3:26-cv-01996-RFL. The two-page ECF 253 identifies the summary-judgment order, final-relief order and final judgment—ECF 250, 251 and 252—as the decisions appealed to the Ninth Circuit. The filing initiates appellate review; it does not itself reverse the judgment or grant a stay. The separate D.C. Circuit procurement-exclusion proceeding is distinct.

Reporting scope: The Meta judgment is dated October 21 and is the Facebook privacy proceeding, separate from the child-safety case. Hansard and the New York court release were reread in complete relevant retained public-source captures after current retrieval failed. The underlying Outlier complaints, Grok complaint and seven-page New York guide were not separately reviewed; those accounts are limited to the linked reporting and official announcement. No later Meta docket status is asserted.