LinkedIn Terms Update: Your Data to Be Used for AI Training by Default (www.linkedin.com)

LinkedIn announced a Terms of Service and Privacy Policy update effective November 3, 2025, clarifying how it will use and share member data across its Affiliates (including Microsoft and its subsidiaries). The update explicitly states LinkedIn may process member data to train the “content‑generating AI models” used in its products and, in some regions, will expand the types of data shared with Microsoft for ad services. EU/EEA/UK/Swiss members are pointed to a regional notice that details the legal basis for model‑training processing, and LinkedIn says members in the affected regions will be offered opt‑out controls. The User Agreement changes also tighten definitions (e.g., “Affiliate”), add exceptions to prior‑notice requirements, clarify appeals processes, ban deepfakes/impersonation, and note secondary payment uses. This matters to the AI/ML community because it signals broader corporate access to high‑quality professional data for model development — potentially improving product personalization and capability while raising questions about consent, provenance, bias, and regulatory compliance. Region‑specific legal bases and opt‑out options mean data availability will vary geographically, affecting reproducibility and dataset curation for researchers and startups. Practitioners should review the regional privacy notices, honor existing user settings (LinkedIn says prior choices remain in force), and factor evolving consent and sharing rules into dataset audits, model governance, and privacy‑compliance processes.