AI Law Wiki News for October 30, 2026 — Episode 29
Ep. 29

AI Law Wiki News for October 30, 2026 — Episode 29

•

Episode description

AI Law Wiki News — October 30, 2026 — Episode 29

Nine first-publication stories in seven discussions: Meta’s content terms; website AI-training preferences; competing copyright arguments; proposed Indian telecom safeguards; SEC fraud allegations; GM driver-data consent; and Work Hub transparency.

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 30 news archive

Source record 1: UK competition regulator investigates Meta’s content-use terms

The UK Competition and Markets Authority opened a Competition Act 1998 investigation on October 30 into Meta’s commercial terms. It will assess market dominance and whether requiring businesses to accept content-use terms for Meta services, including AI features, imposes unfair trading conditions. The CMA says Meta paused the rollout on October 28 following an earlier delay to November 20. No infringement finding has been made. A separate inquiry about Meta One subscriptions and external links is outside this investigation.

Source record 2: Google, OpenAI and Microsoft back shared AI-training preferences

Google, OpenAI and Microsoft announced on October 30 that they plan to honour a shared robots.txt signal for generative-AI training preferences. OpenAI targets the coming months and says existing crawler controls will continue; Google targets 2027 and says explicit Google-Extended choices will remain available. Microsoft says its new training and separate grounding preferences start taking effect as pages are crawled from February 1, 2027, and that NoArchive and NoCache will cease to express those preferences after March 1. This is a voluntary industry initiative with future implementation, not a new copyright exception or an already universal opt-out.

Source record 3: Wikimedia urges different fair-use treatment for AI training and copied answers

The Wikimedia Foundation announced its proposed amicus brief on October 30 in the consolidated New York OpenAI copyright litigation. Its October 29 letter motion, ECF 2361, requests permission to file the proposed brief in support of neither party. Wikimedia argues that transformative model training should qualify as fair use, while retrieval-based answers that substitute for original works should respect publishers’ licensing requirements, including Wikipedia’s attribution and share-alike conditions. These are Wikimedia’s legal arguments and a request for leave, not a court ruling accepting the brief or deciding fair use.

Source record 4: Publisher groups urge rejection of OpenAI’s fair-use defence

The Association of American Publishers, STM, the Independent Book Publishers Association and Copyright Clearance Center announced an October 29 amicus filing in the Southern District of New York’s consolidated OpenAI copyright litigation. The groups ask the court to grant the plaintiffs’ summary-judgment motion and reject the defendants’ position, arguing that existing licensing markets and harms to human authorship weigh against fair use. These are the amici’s arguments, not the court’s findings or a decision on the pending motions.

Source record 5: Authors Alliance challenges market-dilution theory in OpenAI case

Authors Alliance announced on October 30 that it asked for leave the previous day to file an amicus brief supporting the defendants’ summary-judgment motions in the consolidated OpenAI copyright litigation. Its argument focuses on market dilution: it says competition from new works that do not copy protected expression should not count as copyright market harm, and that the theory could burden scholars and smaller authors who rely on fair use. The organisation takes no position on data acquisition or whether particular outputs infringe. This is its proposed amicus position; no order granting leave or resolving the issue is reported.

Source record 6: India’s telecom regulator proposes AI governance and incident-reporting duties

India’s Telecom Regulatory Authority published draft responsible-AI guidelines on October 30 for authorised or licensed entities using or planning AI in telecom networks or services. The proposal calls for risk classification, board-approved internal policies, inventories, supplier due diligence, proportionate human oversight and governance committees led by senior executives. Operators would remain accountable for internally developed and third-party systems and report critical- and high-severity incidents in a prescribed form and timeframe. Comments are invited by November 27; these are consultation proposals rather than operative new duties.

Source record 7: SEC alleges $130 million fraud promoted with AI trading signals

The SEC announced October 29 charges against BG Wealth Sharing Ltd and DSJ Exchange Pty Ltd over a scheme it alleges raised more than $130 million from tens of thousands of retail investors. Its release says the defendants promised extraordinary risk-free crypto returns from purported AI-generated signals and used false SEC filings and certificates to suggest regulatory approval. The Northern District of California complaint seeks injunctions, disgorgement with interest and civil penalties. These are SEC allegations, not a judgment or confirmation that the claimed AI system worked.

Source record 8: GM agrees to $220 million driver-data settlement

Oregon’s attorney general announced on October 29 that a coalition co-led with Washington reached a $220 million settlement with General Motors over allegations that it sold identifiable driving-behaviour and location data to brokers for insurance products without adequate consent. The agency says insurers sometimes used the data to change premiums. The settlement requires affirmative express consent before collection or sharing, limits on collection, a comprehensive privacy programme, internal and external assessments, and dealership oversight and training. This is the new multistate resolution, distinct from earlier FTC and California actions.

Source record 9: DWP discloses AI employment-advice safeguards and review limits

The UK published an algorithmic transparency record for DWP’s Work Hub on October 30. DWP describes a voluntary Jobs and Careers Service tool, available to several thousand private-beta users, that uses Amazon Bedrock and Claude Sonnet 4.6 for CV support, skills analysis, job suggestions and advice. It says the tool does not decide benefit entitlement or automate statutory decisions, and that individual interactions do not receive mandatory human review. The record describes completed May privacy and equality assessments, bias testing and transient external processing. These are departmental disclosures, not independent validation or evidence of a new nationwide launch.

Reporting scope: The Meta inquiry follows the complete CMA release. Training preferences follow all three companies’ announcements and Microsoft’s implementation guide; rollout dates differ and standards work continues. Copyright positions follow complete organizational announcements; Wikimedia’s October29 leave-request letter was independently reread in a full three-page capture saved earlier today because the fresh document route was unavailable. The proposed Wikimedia attachment and separate AAP/Authors Alliance briefs were not independently read in full. These are advocacy positions and requests, not fair-use rulings. India’s full consultation draft was reviewed. The SEC account follows its complete release. Complaint paragraph 23, independently checked in saved PDF text, expressly alleges a proprietary AI data collection system; the full complaint was not independently reread for this episode. GM settlement terms follow the complete Oregon announcement, not separate consent-judgment review. Work Hub safeguards are departmental representations. Substantive case pages provide background; their archive cross-links do not independently substantiate new filings.

0:00Opening
0:37Britain investigates Meta's content terms
2:00A shared training signal, with different timetables
3:48Competing copyright arguments in New York
5:33India's proposed telecom AI safeguards
6:43SEC alleges a crypto-investment fraud
7:56GM's driver-data settlement
9:05Britain's Work Hub transparency record
10:13Closing