Notion: customer AI chat transcripts exposed via Fireworks AI breach (LLMJacking-linked)
Severity: medium · Patch status: no-fix-needed · Since 2026-11-12
Fireworks AI, a cloud platform for running and fine-tuning open-source AI models, was hacked twice in October 2026: once via compromised internal credentials and once on October 24 through a third-party cloud provider hosting Fireworks R&D machines, in an incident linked to rising LLMJacking activity. Attackers executed an unauthorized workload that compromised customer data. Fireworks contained the incident the same day and notified Notion on November 2. Notion's investigation found attackers accessed Notion AI chat transcripts used for the Insights feature; fewer than 1% of workspaces were affected and most transcripts were unattributable to any customer or workspace. Notion has notified affected customers, urging them to treat the data as exposed and rotate any credentials pasted into Notion AI during the incident window.
What to do
- If you use Notion AI, check for a breach notification and securely review your workspace data as instructed
- Rotate any passwords, API keys, or credentials pasted into Notion AI during the October incident window
- Review Notion AI chat history for sensitive data and adjust usage policies to avoid pasting secrets into AI features
- If you are a Fireworks AI customer, review inference logs and API usage for unauthorized workloads
- Monitor for follow-on phishing or extortion attempts referencing exposed chat content
Sources
All live incidents on Threat Watch
