Last Update: 09/29/2026 at 4:33 PM EST

Morning Briefing: Privacy

Saturday, September 19, 2026

September 19, 2026

Privacy Rules Are Being Built One Sector at a Time

The clearest privacy developments pointed in different directions, but toward the same practical reality: safeguards are being negotiated and contested in specific settings rather than arriving through one uniform framework. A Senate hearing will examine the reach of networked vehicle surveillance, while Microsoft’s school AI agreement puts binding limits on secondary uses of student data.

California’s pending rollback of a narrow class of tracking suits and Revolut’s confirmed disclosure through a fake government request sharpen the contrast. Legal exposure may shift through legislation, but sensitive data can still be lost through ordinary verification failures.

Federal scrutiny of Flock Safety’s vehicle-surveillance network has moved beyond requests for information to a September 23 Senate Judiciary subcommittee hearing. Fox News reported that Flock, Axon, Motorola Solutions and Verkada executives were invited. The hearing could bring retention, cross-jurisdictional sharing, access controls and warrantless-search practices into clearer public view, though it creates no new rule by itself.

Microsoft and the American Federation of Teachers have set a binding, November 1 framework for covered schools that generally bars use of student and educator data for AI training, advertising, sales and unrelated product development. Audits, family-facing explanations and limits on emotionally dependent AI features make this more operationally meaningful than a broad statement of principle.

California’s SB 690 is now awaiting Governor Gavin Newsom’s decision. If enacted, it would reserve certain CIPA Section 638.51 claims tied to website and app tracking to the state attorney general, potentially narrowing a major source of private litigation. It would not settle the underlying tracking question or remove other privacy, consent and vendor-governance obligations.

Revolut confirmed that staff disclosed sensitive customer information after receiving a fraudulent request from an address on a legitimate government domain. The company says its systems and customer funds were unaffected, but the possible exposure of identity documents, account records and transaction histories shows how a trusted-channel impersonation can create serious privacy harm without a conventional systems breach.

Key Points

  • Surveillance oversight is becoming more concrete, but remains preliminary. The Flock hearing extends recent congressional pressure into a forum that may expose operating practices; it does not yet establish federal limits on networked vehicle-location searches.
  • Contract terms are doing privacy work that broader rules have not yet standardized. Microsoft’s school agreement offers districts a usable model for data-use restrictions, audits and product-design limits, but its reach remains confined to covered contracts.
  • The day also underscores that data governance is not only a question of what products collect. Revolut’s incident puts equal weight on how employees authenticate and escalate exceptional requests for already-held sensitive data.

Implications

School districts and education-technology buyers now have a more concrete benchmark for negotiating AI data-use, audit and disclosure terms. Its wider influence will depend on implementation and whether competing vendors adopt comparable commitments.

Surveillance vendors and their public-sector customers should expect closer attention to searchable data sharing, retention and access logs. The immediate test is what executives disclose at the hearing, not an assumed legislative outcome.

SB 690 could reduce one California tracking-litigation pathway if it takes effect, but businesses would still need to assess tracking disclosures, consent, vendor data flows and exposure under other legal theories.

Financial institutions and other sensitive-data custodians have a clear operational reminder: government and law-enforcement request channels require strong authentication and escalation controls, not merely secure core infrastructure.

Watchpoints

Watch

Whether invited Flock and other surveillance-technology executives testify on September 23, and whether the hearing leads to further oversight or legislative action.

Watch

Governor Newsom’s decision on SB 690 by September 30, including any subsequent challenge to its proposed retroactive effect.

Watch

Revolut’s disclosures on the affected population, notification scope and any regulator response.

Watch

Whether other major AI vendors adopt binding student-data restrictions comparable to Microsoft’s agreement.

Fallout

The day’s developments show privacy governance advancing through targeted contracts, pending state legislation and oversight, while disclosure workflows remain a direct source of exposure.

Networked Vehicle Surveillance

Flock’s camera and vehicle-location network is moving into higher-profile federal oversight.

Fresh developments

A Senate Judiciary subcommittee scheduled a September 23 hearing on access, retention, sharing, cybersecurity and warrantless-search concerns, with several surveillance-technology executives invited.

Why we noticed

Recent scrutiny has now become a formal hearing, increasing pressure for clearer answers on how networked location data is governed without yet creating new legal constraints.

Watch for:

  • Confirmed witness participation and testimony.
  • Specific disclosures on retention, access controls and interagency sharing.
  • Any follow-on oversight or legislative action.

Student AI Data Governance

Microsoft’s agreement with the American Federation of Teachers creates binding limits for covered schools’ use of AI tools.

Fresh developments

The agreement, effective November 1, generally restricts use of student and educator data for AI training, advertising, sales and unrelated product development, while requiring audits and family-facing explanations.

Why we noticed

The terms offer a practical procurement model for education AI, including both data-use limits and restrictions on features designed to foster emotional dependency.

Watch for:

  • How the agreement is implemented and audited.
  • Whether its coverage and enforcement mechanics become clearer.
  • Whether other major AI vendors make comparable binding commitments.

California Tracking Litigation

SB 690 could narrow private litigation over a specific California wiretapping-law theory applied to online tracking.

Fresh developments

The unanimously passed bill awaits Governor Newsom’s action and would place qualifying CIPA Section 638.51 claims involving websites and apps in the attorney general’s hands rather than private plaintiffs’.

Why we noticed

The measure could materially change litigation exposure around pixels, cookies, session replay and similar tools, while leaving other privacy and tracking obligations intact.

Watch for:

  • Governor Newsom’s action by September 30.
  • Whether the bill takes effect on January 1, 2027.
  • Litigation over the bill’s retroactive reach and remaining CIPA claims.

Sensitive Financial-Data Disclosure Controls

Revolut’s confirmed incident illustrates the privacy risk in trusted-request workflows.

Fresh developments

A fraudulent request sent from a legitimate government-domain address led staff to disclose sensitive customer data to an unauthorized party; Revolut says its core systems and customer funds were unaffected.

Why we noticed

Identity documents, verification selfies, account statements and transaction histories can enable targeted fraud even when no core system is compromised.

Watch for:

  • The number and location of affected customers.
  • The scope of notifications and regulator engagement.
  • Any further confirmed detail on the disclosure and request-validation controls.

Final Thought

Privacy protection is becoming more operationally specific—written into contracts, tested in oversight hearings and reshaped by state law—yet the weakest point can still be the human process governing access to data already collected.