> ## Content Index
> Fetch the complete content index at: https://aigovernancebrief.org/llms.txt
> Use this file to discover other available public pages before exploring further.

# Weekly AI Governance Brief: 9–15 November 2026
- URL: https://aigovernancebrief.org/weekly-ai-governance-brief-9-15-november-2026/
- Published: 2026-11-16T16:38:57.000Z
- Updated: 2026-11-16T16:38:57.000Z
- Author: AI Governance Brief Team

Bringing you the latest developments in the AI Governance world.

## **European Commission proposes Digital Fairness Act**

On 11 November, the European Commission adopted its proposal for a Digital Fairness Act, accompanied by a Simplification Directive and proposed changes to the Consumer Protection Cooperation Regulation. The package seeks to update EU consumer protection rules for digital services and address commercial practices that the Commission considers insufficiently covered by existing legislation.

The proposed Digital Fairness Act targets manipulative interface design, including misleading scarcity claims that pressure consumers into purchases. It also addresses addictive design features that encourage excessive engagement or spending, unfair personalisation targeting consumer vulnerabilities and opaque dynamic pricing practices. Other provisions concern influencer advertising and obstacles to terminating digital subscriptions.

The Commission proposes additional protections for minors, including restrictions on unfair personalised advertising and pricing directed at children. It also seeks to improve transparency in online purchasing, including clearer information about pricing in virtual queues and stronger safeguards against unwanted subscription renewals.

The proposal forms part of a wider effort to modernise consumer protection enforcement. It remains subject to the EU legislative process and does not yet establish binding new obligations for digital service providers.

## **EDPB selects AI-driven recruitment for coordinated enforcement**

On 11 November, the European Data Protection Board announced that automated decision-making in recruitment, including AI-driven candidate screening, will be the subject of its sixth Coordinated Enforcement Framework action. The initiative is scheduled to begin during 2027, with national data protection authorities participating voluntarily.

The coordinated action will examine compliance with GDPR protections concerning decisions based solely on automated processing. It will also address the associated right of access, particularly where recruitment systems evaluate or screen candidates through automated tools.

The EDPB's selection is significant because it identifies a specific application of AI for coordinated supervisory attention. Rather than focusing on AI systems generally, the exercise will examine how automated recruitment decisions interact with existing data protection rights.

The announcement does not introduce new GDPR requirements. It establishes a formal enforcement priority and the framework through which participating national authorities will examine relevant practices.

European Data Protection Board

### **Why this matters**

AI-assisted recruitment involves decisions that can directly affect individuals' access to employment. The EDPB's initiative places attention on the circumstances in which automated screening may fall within GDPR restrictions on solely automated decision-making and how individuals can exercise their information rights.

The coordinated approach also has institutional significance. Participating national authorities will examine a common subject under the EDPB's framework, allowing supervisory findings to be considered across jurisdictions.

For organisations using automated candidate-screening tools, the development identifies a concrete area of future regulatory scrutiny. It does not establish that all AI-assisted recruitment is prohibited or that every automated screening process falls within the same legal category.

## **EU AI Act product provisions incorporated into the Windsor Framework**

On 10 November, the EU–UK Withdrawal Agreement Joint Committee adopted Decision No 3/2026, adding specified provisions of the EU Artificial Intelligence Act to Annex 2 of the Windsor Framework. The decision entered into force on 11 November and is binding on the European Union and the United Kingdom under the Withdrawal Agreement.

The decision concerns provisions of the AI Act relating to conditions and technical specifications for placing products on the market or putting them into service, insofar as the free movement of products is concerned. It expressly excludes Articles 102, 108 and 110 from the addition.

The legal mechanism is narrower than a general extension of the AI Act throughout the United Kingdom. It concerns the application of relevant product-related EU rules under Northern Ireland's arrangements within the Windsor Framework.

The decision also distinguishes provisions already applicable through Article 13(3) of the Windsor Framework from those newly added to Annex 2\. It therefore represents a formal change in the legal framework governing relevant AI-enabled products rather than a general UK policy announcement.

## **UK ICO clarifies data protection expectations for agentic AI advertising**

On 13 November, the UK Information Commissioner's Office published *AdTech 3.0: Online advertising in the age of LLMs and agentic AI*. The report examines the growing use of large language models and autonomous agents in advertising activities, including campaign planning, audience selection and personalised advertising delivery.

The ICO's central conclusion is that these technologies do not create a separate set of data protection requirements. However, their capabilities can change how organisations must understand and apply existing obligations under UK data protection legislation and the Privacy and Electronic Communications Regulations.

The report places particular emphasis on accountability. Organisations remain responsible for personal data processing even when autonomous agents perform operational activities or optimise advertising campaigns. The use of an AI system does not transfer that responsibility away from the organisation determining the purposes and means of processing.

The ICO also addresses the potential for AI systems to generate inferences from conversational inputs and behavioural information. Such inferences may reveal information that individuals did not explicitly provide, including potentially sensitive characteristics. The report therefore examines how profiling, transparency and user choice apply when advertising is delivered through conversational interfaces.

Another concern is the increasing dependence on technology suppliers. The ICO states that organisations need sufficient evidence concerning data sources, processing purposes and operational controls, rather than relying exclusively on contractual assurances. The report applies existing law to emerging practices and does not create new legal requirements.

## **UK consults on AI product safety certification for education**

On 10 November, the UK Department for Education opened a consultation on introducing a certification scheme for education technology and AI products used in schools and other educational settings in England. The consultation will remain open until 4 February 2027.

The proposed scheme is intended to provide greater assurance about the safety of products supplied to educational institutions. It would offer suppliers a route to demonstrate compliance with defined safety standards while helping schools and other education providers make more informed procurement decisions.

The consultation also examines how a certification framework could reduce the assessment burden on individual educational institutions. A central consideration is whether product safety assurance can be organised in a way that is proportionate for suppliers while providing meaningful information to organisations purchasing or deploying AI-enabled educational tools.

The proposal remains at the consultation stage. The Department has not established a final certification framework or introduced new mandatory requirements through this publication.

## **NIST publishes draft adversarial machine-learning taxonomy for agentic AI**

On 9 November, the US National Institute of Standards and Technology released an initial public draft of *Adversarial Machine Learning: A Taxonomy and Terminology of Attacks and Mitigations*, designated NIST SP 2400-102e2026\. The document is open for public comment until 18 December.

The draft focuses on attacks against agentic AI systems and the methods available to mitigate their consequences. It establishes a structured taxonomy covering machine-learning methods, stages in the attack lifecycle and characteristics of attackers, including their objectives, capabilities and knowledge.

NIST also introduces a dedicated index for agentic AI attacks, allowing more precise referencing of specific threat categories. The publication draws on technical research and engagement with experts in adversarial machine learning.

The stated purpose is to develop a common vocabulary that can support future standards and practical guidance concerning the security of increasingly autonomous AI systems. The draft is not a binding regulatory instrument and remains subject to consultation.

## **Looking ahead**

The developments of 9–15 November demonstrate several distinct approaches to AI governance. At EU level, the Commission has proposed new consumer protection rules addressing digital personalisation, while the EDPB has selected automated recruitment for coordinated enforcement. The EU–UK Joint Committee has also taken a binding step concerning the territorial application of product-related AI Act provisions.

Outside the EU, the ICO has applied existing data protection requirements to agentic advertising, the Department for Education has proposed a product safety certification mechanism and NIST has released a technical taxonomy for adversarial AI risks. Across these developments, the emphasis is on defining the responsibilities of organisations deploying AI and establishing more specific mechanisms for evaluating automated systems.

## **Sources**

European Commission, Digital Fairness Act proposal and accompanying legislative documents: [https://commission.europa.eu/law/law-topic/consumer-protection-law/digital-fairness-act\_en](https://commission.europa.eu/law/law-topic/consumer-protection-law/digital-fairness-act%5Fen?utm%5Fsource=chatgpt.com)

European Data Protection Board, plenary decisions and 2027 coordinated enforcement topic: [https://www.edpb.europa.eu/news/edpb-adopts-statement-on-review-of-the-republic-of-korea-adequacy-processor-bcr\_en](https://www.edpb.europa.eu/news/edpb-adopts-statement-on-review-of-the-republic-of-korea-adequacy-processor-bcr%5Fen?utm%5Fsource=chatgpt.com)

EU–UK Joint Committee, Decision No 3/2026 concerning the AI Act and Windsor Framework: [https://www.gov.uk/government/publications/decision-no-32026-of-the-joint-committee-established-by-the-agreement-on-the-withdrawal-of-the-uk-from-the-eu-adding-a-newly-adopted-union-act-to-anne/decision-no-32026-of-the-joint-committee-established-by-the-agreement-on-the-withdrawal-of-the-uk-from-the-eu-adding-a-newly-adopted-union-act-to-an](https://www.gov.uk/government/publications/decision-no-32026-of-the-joint-committee-established-by-the-agreement-on-the-withdrawal-of-the-uk-from-the-eu-adding-a-newly-adopted-union-act-to-anne/decision-no-32026-of-the-joint-committee-established-by-the-agreement-on-the-withdrawal-of-the-uk-from-the-eu-adding-a-newly-adopted-union-act-to-an?utm%5Fsource=chatgpt.com)

UK Information Commissioner's Office, *AdTech 3.0: Online advertising in the age of LLMs and agentic AI*: [https://ico.org.uk/media2/nt4jmexg/adtech-3-0-202611.pdf](https://ico.org.uk/media2/nt4jmexg/adtech-3-0-202611.pdf?utm%5Fsource=chatgpt.com)

UK Department for Education, consultation on EdTech and AI product safety certification: [https://www.gov.uk/government/consultations/introducing-an-edtech-and-ai-product-safety-certification-scheme](https://www.gov.uk/government/consultations/introducing-an-edtech-and-ai-product-safety-certification-scheme?utm%5Fsource=chatgpt.com)

US National Institute of Standards and Technology, draft SP 2400-102e2026 on adversarial machine learning: [https://www.nist.gov/news-events/news/2026/11/adversarial-machine-learning-taxonomy-and-terminology-draft-nist-sp-2400](https://www.nist.gov/news-events/news/2026/11/adversarial-machine-learning-taxonomy-and-terminology-draft-nist-sp-2400?utm%5Fsource=chatgpt.com)