• Commission: preliminary DSA breach findings for TikTok & Meta. The Commission signalled non-compliance with DSA transparency obligations tied to researchers’ access to public data, key for accountability research. (EC — Digital Strategy press release, 24 Oct 2025). Digital Strategy

  • UK unveils dementia R&D challenge including AI diagnostics. The government announced targeted funding to support AI-enabled earlier diagnosis and treatment pathways, with explicit 2029 targets. (GOV.UK press release, 24 Oct 2025). GOV.UK

  • EU AI Board meets (5th meeting). The Board convened today to coordinate AI Act rollout, including work on guidance and interfaces with sector rules. (EU AI Board page). Digital Strategy


Regulation

  • Researcher-access enforcement under the DSA. The Commission’s preliminary view against TikTok/Meta underscores that data-access for vetted researchers is an enforceable legal duty, not a courtesy—material for AI accountability studies and risk audits. (EC press note). Digital Strategy

  • UK digital identity governance (register updated guidance). Fresh guidance on the statutory register for digital identity & attribute services strengthens transparency around certification—relevant for AI-enabled identity proofing and liability allocation. (GOV.UK — Smart Data Council co-chair appointment, 24 Oct 2025). GOV.UK


Academia

  • Prajescu & Confalonieri — Argumentation-Based Explainability for Legal AI (arXiv). Proposes argumentation-centric XAI tailored to legal reasoning and forthcoming AI Act/GDPR explainability expectations. (arXiv).

  • Sunny et al. — Explainability and Trust (Quantitative Study) (arXiv). Empirical link between explanation modalities and user trust—useful for risk management and documentation duties. (arXiv).

  • “Label Indeterminacy in AI & Law” (arXiv). Shows how ambiguous training labels skew legal-domain model behaviour—relevant to traceability and dataset governance under the AI Act. (arXiv PDF). 

  • Werner — The Impossible Act: Structural Incompatibilities within EU Law (SSRN). Frames a compliance “trilemma” across copyright/consumer rights/fundamental rights for AI providers. (SSRN). 


Business

  • Compliance by access: Today’s DSA action signals that researcher-data access is part of the accountability stack; platforms and AI providers should harden request workflows, auditing, and disclosures to avoid enforcement risk. 


Adoption of AI

  • Health-AI pipelines must be audit-ready. UK dementia challenge implies DPIAs, evaluation plans, and clinical-safety documentation before deployment. 

  • Governance coordination continues. The AI Board meeting reflects ongoing alignment on guidance and standards needed to operationalise AI Act duties. 


Takeaway

Access and evidence are the day’s themes. DSA enforcement pressures platforms to open data for scrutiny, UK health-AI funding demands verifiable evaluation, and the AI Board’s steady cadence keeps implementation on track. Organisations should treat researcher access, dataset documentation, and explainability as non-negotiable compliance artefacts.


Sources: European Commission, GOV.UK, EU AI Board, Reuters, arXiv, SSRN