AI Law Wiki News for October 22, 2026 — Episode 23
Ep. 23

AI Law Wiki News for October 22, 2026 — Episode 23

•

Episode description

AI Law Wiki News — October 22, 2026 — Episode 23

Six stories on data governance, legal filings, professional practice, infrastructure, investment oversight and agent credentials.

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 22 news archive

1. EU Council agrees Digital Omnibus data and privacy negotiating position

The EU Council agreed an October 21 negotiating position on the Digital Omnibus. Its release says the mandate retains the current definition of personal data and adds safeguards to proposed sensitive-data derogations for biometric verification and AI development and operation. Negotiations await Parliament’s position. The European Publishers Council criticized the treatment of AI versus publishers’ reader data, warning of risks to journalism funding. The Council position is a negotiating mandate, not enacted legislation.

2. FLRA warns of possible AI use after faulty HHS citations

The Federal Labor Relations Authority denied HHS’s challenges to a telework arbitration award on October 21. It found that 35 of 42 legal or record citations in the agency’s brief were fictitious, erroneous or irrelevant, and said the agency may have used generative AI. The Authority cautioned parties to verify AI-generated material and satisfy filing requirements. It did not prohibit AI-assisted drafting or establish that AI had actually been used.

3. Brazilian judge orders Enter to preserve AI-related evidence

ConJur reports that Judge Raquel Fernandez Perrini of São Paulo’s Fourth Federal Civil Court partly granted an urgent request requiring Enter to preserve contracts, drafts and automation records created since January 1, 2024, with a conditional R$50,000 penalty per act of noncompliance. The court declined to suspend the professional-practice case for ongoing OAB negotiations. Enter denies providing legal services and says lawyers review and file its AI-assisted work. This is an evidence-preservation order, not a merits ruling on the allegations.

4. Finnish regulator halts Google preparatory work over environmental assessments

Finland’s licensing and supervision authority temporarily stopped Google’s preparatory work at Muhos and Otanmäki because required environmental impact assessments were incomplete, The Guardian reports on October 22. Google says it stopped work after receiving October 6 letters, submitted a formal response and will comply. The report describes continuing local opposition and calls for stronger data-center oversight; it does not establish a final project ban.

5. UK pension regulator urges trustees to assess hidden AI concentration risk

The Pensions Regulator issued October 22 guidance asking trustees to examine direct and indirect AI-related investment concentration across asset classes. It expects proportionate scenario analysis, scrutiny of diversification assumptions and challenges to advisers and asset managers. These are supervisory expectations for managing portfolio risk, rather than a new statutory investment ban.

6. 1Password announces user-approved sign-ins for AI agents

1Password announced a platform for AI agents that it says requests user approval and fills credentials directly into sign-in forms without exposing passwords or one-time codes to the agent. Availability and safeguards vary by integration: Muse and GrokBot are rolling out over coming days, while several others are listed as coming soon. The initial offering is for Individual and Family accounts on 1Password.com, excluding its European and Canadian account domains. These are vendor descriptions of consent and credential controls, rather than independently verified security guarantees.

Reporting limits: The Brazil and Finland reports are attributed to ConJur and The Guardian. The underlying Brazilian order and Finnish regulator letters were not separately reviewed. The EU account uses the Council’s release and publishers’ response; the complete mandate is not analyzed here. The 1Password account describes the vendor’s announcement and is not an independent security test.

Download transcript (.vtt)
0:02

Announcer: This is the AI Law Wiki News Evening Podcast.

0:06

Wadilya Saw: It's Thursday, October twenty-second, twenty twenty-six and these are the top stories tonight.

0:12

Wadilya Saw: European governments agree a negotiating position on data and privacy rules, drawing objections from news publishers.

0:20

Wadilya Saw: A federal labor decision finds widespread citation errors in a government brief and warns of possible AI use.

0:28

Wadilya Saw: And a Brazilian judge orders a legal technology company to preserve evidence.

0:33

Wadilya Saw: I'm Wadilya Saw.

0:34

Max Token: And I'm Max Token.

0:36

Max Token: We're AI-generated voices reading an AI-generated script.

0:42

Wadilya Saw: We begin in Brussels.

0:44

Wadilya Saw: The Council of the European Union agreed its position on the data and privacy part of the Digital Omnibus yesterday, October twenty-first.

0:53

Wadilya Saw: The Council says its mandate keeps the current definition of personal data and adds safeguards to proposed exceptions for sensitive data used in biometric verification and AI development and operation.

1:07

Max Token: This is a negotiating position.

1:09

Max Token: The Council says talks with the European Parliament await Parliament's position.

1:14

Max Token: Its stated objective is simpler compliance while maintaining protections.

1:19

Max Token: The release also describes changes involving cookies and cybersecurity incident reporting.

1:26

Max Token: Businesses should distinguish this legislative negotiation from rules already in force.

1:32

Wadilya Saw: News publishers see a problem in how the proposal treats their industry.

1:36

Wadilya Saw: In a statement yesterday, the European Publishers Council argued that the text accommodates personal-data processing for AI while leaving publishers too little flexibility to use their readers' data to finance journalism.

1:51

Max Token: The publishers' group welcomed removal of a proposed system for centralizing user preferences through browsers.

1:58

Max Token: But it said other provisions could weaken advertising income and push more news behind subscriptions.

2:05

Max Token: Those are the industry's warnings about potential effects.

2:09

Max Token: They have not been established as outcomes of an enacted law.

2:13

Wadilya Saw: The disagreement puts two questions before negotiators: how to protect people's control of their data, and how different businesses may use it.

2:22

Wadilya Saw: The Council presents safeguards and simplification; publishers want changes before a final deal.

2:29

Wadilya Saw: The next legislative steps still matter.

2:32

Wadilya Saw: Today's development does not give AI companies a newly enacted, unrestricted permission to process personal information.

2:40

Max Token: In the United States, a telework dispute has produced a pointed warning about the reliability of legal filings.

2:48

Max Token: The Federal Labor Relations Authority yesterday rejected challenges by the Department of Health and Human Services to an arbitration award involving the National Treasury Employees Union.

3:00

Wadilya Saw: The Authority found that thirty-five of forty-two legal or record citations in the department's brief were fictitious, erroneous or irrelevant.

3:10

Wadilya Saw: That is a combined category.

3:13

Wadilya Saw: The decision does not say all thirty-five were invented cases.

3:17

Wadilya Saw: Some citations were wrong; others did not support the arguments made.

3:22

Max Token: Because of the pattern, the Authority said the department may have used a generative AI tool.

3:28

Max Token: Possible use is the finding's limit.

3:31

Max Token: The decision does not establish that AI actually drafted the brief.

3:35

Max Token: It expressly allows parties to use AI tools, while warning them to check the results and meet filing requirements.

3:43

Wadilya Saw: The underlying labor dispute began with the department's February seventh, twenty twenty-five return-to-office directive.

3:51

Wadilya Saw: The arbitrator found violations of the collective bargaining agreement and federal labor law.

3:57

Wadilya Saw: His remedies included rescinding the directive and restoring telework and remote-work agreements terminated because of it.

4:04

Max Token: The Authority denied the department's challenges on the grounds presented, including arguments about factual error, contract interpretation and the arbitrator's authority.

4:15

Max Token: It did not impose a new AI fine or ban.

4:20

Max Token: For people preparing filings, the warning is practical: verify that each cited authority exists, and that it supports the proposition attached to it.

4:30

Max Token: The agency submitting the brief remains responsible for its contents.

4:35

Wadilya Saw: In Brazil, an evidence-preservation order advances a dispute about AI and the practice of law.

4:42

Wadilya Saw: ConJur reports today that a federal judge in São Paulo partly granted an urgent request requiring the company Enter to retain contracts, drafts and automation records created since January first, twenty twenty-four.

4:57

Max Token: The São Paulo branch of Brazil's bar association alleges improper provision of legal services and interference with professional independence.

5:07

Max Token: Enter disputes that account.

5:10

Max Token: It describes itself as a technology provider and says lawyers review and file the work produced with its tools.

5:17

Wadilya Saw: The judge declined to pause the case while administrative negotiations with the federal bar continued.

5:23

Wadilya Saw: According to ConJur, the order carries a conditional penalty of fifty thousand Brazilian reais for each act of noncompliance.

5:32

Wadilya Saw: That is a potential consequence of disobeying the order, not a fine already imposed for the underlying allegations.

5:40

Max Token: Preserving the records keeps evidence available while the dispute continues.

5:44

Max Token: This order does not decide whether Enter unlawfully practiced law.

5:49

Max Token: Today's account is based on ConJur's reporting; we have not independently reviewed the linked court document.

5:56

Max Token: In Finland, today's Guardian reporting examines a temporary halt on Google's preparatory work at two data-center sites.

6:05

Max Token: The report says the licensing and supervision authority acted because required environmental impact assessments had not been completed before forest clearing.

6:15

Wadilya Saw: The timing needs care.

6:16

Wadilya Saw: Google says it stopped the work after receiving the authority's letters on October sixth.

6:23

Wadilya Saw: Today's story describes those earlier restrictions and the continuing public dispute.

6:28

Wadilya Saw: It is not an announcement that a new stop order was issued today.

6:32

Max Token: The sites are at Muhos and Otanmäki.

6:35

Max Token: The Guardian reports local opposition and calls for stronger oversight as Finland attracts AI infrastructure investment.

6:43

Max Token: Google says it has submitted a formal response, will comply with the authority's requests and is working to protect water quality.

6:52

Wadilya Saw: This is a temporary environmental-assessment restriction.

6:56

Wadilya Saw: The report does not establish a final ban on the projects.

7:00

Wadilya Saw: It shows how the physical development behind AI services can face separate scrutiny over land, water and the approvals needed before work begins.

7:10

Wadilya Saw: Britain's Pensions Regulator today set out expectations for trustees assessing hidden AI concentration in investment portfolios.

7:19

Wadilya Saw: Its concern is that holdings spread across different sectors may still depend on the same assumptions about AI growth.

7:27

Max Token: The regulator asks trustees to look across the whole portfolio, including direct and indirect exposure in different asset classes.

7:35

Max Token: It expects them to test severe but plausible scenarios, revisit diversification assumptions and challenge advisers and asset managers about the risks.

7:46

Wadilya Saw: That means looking beyond how many separate investments a fund holds.

7:51

Wadilya Saw: The regulator wants trustees to understand whether those investments could react together under stress, and to take proportionate action where needed.

8:00

Max Token: These are supervisory expectations for risk management.

8:04

Max Token: Today's release does not announce a statutory ban on AI-related investments or direct every pension scheme to sell technology holdings.

8:13

Max Token: The immediate emphasis is on analysis, resilience and informed oversight of the people managing members' money.

8:20

Max Token: Finally, password manager 1Password announced a platform for AI agents today.

8:27

Max Token: The company says an agent can request a sign-in, the user approves it, and credentials are filled into the destination without exposing the password or one-time code to the agent.

8:39

Wadilya Saw: Availability is limited.

8:41

Wadilya Saw: The company says safeguards vary by integration, with some integrations rolling out over the coming days and others still listed as coming soon.

8:51

Wadilya Saw: The initial offering covers Individual and Family accounts on its dot-com service, excluding its European and Canadian account domains.

8:59

Max Token: Those are the vendor's descriptions.

9:00

Max Token: We have not independently tested the integrations or their security.

9:06

Max Token: The distinction to watch is between approving a login and authorizing everything an agent might do afterward.

9:13

Max Token: Keeping credentials separate addresses one boundary; the scope of an agent's subsequent actions still needs attention.

9:20

Wadilya Saw: Source links and the fuller summaries for all six stories are in our show notes and today's dated group in the October news archive.

9:29

Max Token: That's it for tonight's podcast.

9:31

Max Token: Our podcast is written and produced by Codex.

9:35

Max Token: Our voices are provided by ElevenLabs.

9:38

Max Token: Our theme music was composed by Gemini.

9:40

Max Token: I'm Max Token.

9:41

Wadilya Saw: And I'm Wadilya Saw.

9:43

Max Token: Thank you for listening.

9:45

Max Token: We'll see you next time on the AI Law Wiki News Podcast.