Notion disclosed a security incident involving its third-party AI infrastructure vendor, Fireworks AI, which may have exposed unattributable Notion AI chat transcripts from fewer than 1% of Notion workspaces. Fireworks, a platform serving tens of thousands of customers and hundreds of thousands of developers, was used by Notion for classification jobs powering the Insights feature and as a general inference provider for open-weight models. Notion learned of potential incidents earlier in November, severed all connections to Fireworks on November 3, and deleted all data from Fireworks systems. On November 6, Fireworks notified Notion that malicious actors may have accessed customer data, and confirmed a specific dataset on November 8. Analysis revealed this dataset contained unattributable chat transcripts used for Insights; these transcripts lacked user or workspace identifiers and did not include tool call results such as page content or external data. No other workspace data was affected. Notion has disabled the Insights feature and ceased serving open-weight models via Fireworks. Affected workspace owners have been contacted and can review exposed data via the workspace sidebar under Security incident review (desktop browser only). Enterprise plan owners receive CSV and annotated CSV exports with metadata for internal analysis; other plans receive a shareable link for members to access their own data. Notion advises treating the data as exposed and rotating any credentials pasted into Notion AI during the relevant timeframe.