🔍 Read the full analysis: Notion Reports Fireworks Breach Exposure Of AI Chat Transcripts on ThorstenMeyerAI.com
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
Notion says a dataset accessed in the Fireworks AI breach included stored Notion AI chat transcripts used for its Insights feature, affecting fewer than 1% of workspaces. The company says it revoked exposed Notion API keys found in affected messages on November 11, but has not identified which Fireworks system retained the data or for how long.
Notion says chat transcripts from its AI product were included in a dataset accessed during the Fireworks AI breach, expanding its account of the incident and raising questions about data-retention commitments. The company estimates that fewer than 1% of Notion workspaces were affected and says it revoked exposed Notion API keys found in affected user messages on November 11.
In its updated security advisory and Help Center notice, Notion said the dataset Fireworks provided on November 8 included Notion AI chat transcripts used for its Insights feature. Insights gives workspace administrators information about how teams use AI. Notion said the transcripts in most affected workspaces were “unattributable”: they were not labeled with the author or workspace and did not include tool-call results. That does not mean the text was necessarily anonymous in practice or free of sensitive content.
Notion defines a transcript as any input or output written to the chat. It said that material could contain sensitive information and advised affected customers to review it. The company said AI Meeting Notes were not affected unless audio, transcripts or summaries were used as input in a Notion AI chat. It also said no other Notion features were affected.
Notion said the breach affected fewer than 1% of workspaces. It advised customers who receive an incident notice to treat the relevant data as exposed, review it under its Security incident review guidance and rotate passwords, API keys or other credentials pasted into Notion AI during the incident. On November 11, Notion said it had revoked exposed Notion API keys found in affected messages. The company’s notice does not establish how many individual users or messages were involved.
Notion Reports Fireworks Breach Exposure of AI Chat Transcripts
A dataset accessed in the Fireworks AI breach included stored Notion AI chat transcripts used for Insights. The exposure highlights the gap between live model requests and data retained to power product features.
What the dataset contained
Notion says the exposed material came from Insights classification work, not its general inference service.
AI chat transcripts
Inputs and outputs written to Notion AI chats could contain sensitive information.
Other features
Notion says no other features were affected. AI Meeting Notes were included only if their content was used in a Notion AI chat.
How the incident unfolded
What customers should review
Notion says affected organizations should treat relevant data as exposed.
- Review potentially exposed Notion AI chat inputs and outputs.
- Rotate passwords, API keys or other credentials pasted into chats during the incident.
- Follow Notion’s Security incident review guidance if your workspace received a notice.
Retention questions remain
Notion says its contractual zero-retention commitment covered the Fireworks analysis, yet data from the endpoints it used was retained.
- Most affected transcripts lacked author and workspace labels.
- It is auditing critical vendors for data-minimization practices.
- Which Fireworks system retained the data, and for how long.
- How many users or transcript entries were involved.
- Whether data remained accessible after Notion disconnected.
Stored Chats Put Data Promises to the Test
The incident matters because AI chat transcripts can contain more than routine prompts. Users may enter customer names, contract details, internal plans or credentials. Notion has specifically warned that affected inputs and outputs may include sensitive material, so the practical risk depends on what people typed and what the records contained.
The case also highlights a distinction between zero data retention for live model requests and data stored to support a product feature. Notion says its zero-retention commitments for model vendors cover API requests used to generate responses, but do not cover stored data needed for features such as chat history and Insights. For Insights, Notion says it sent stored AI requests and responses to Fireworks for analysis under a contractual zero-retention commitment. It later learned that data from the endpoints it used had been retained.
That account leaves customers with a question about how a contractual safeguard interacted with the systems involved. A commitment can set expectations, but the exposure shows that customers still need clear information about what data is stored, where it is processed and how vendors verify deletion or retention limits.
Products Worth Considering
Logitech MX Keys S Combo Wireless Keyboard and Mouse, Graphite
Bose TV Speaker Compact Soundbar with Bluetooth and HDMI-ARC, Black
LG UltraGear 34G600A-B 34-Inch Curved WQHD Gaming Monitor, 160Hz
As an affiliate, we earn on qualifying purchases.
How Notion and Fireworks Used the Data
Notion said it used Fireworks for two purposes: classification jobs for Insights and inference for open-weight models in its product. According to Notion, the exposed dataset came from the Insights work, not the general inference service. Fireworks’ own bulletin, last updated November 9 according to the source material, said it had not detected an impact to Inference, Training SDK and Nexus, which it describes as zero data retention by default, while listing datasets among data accessed for certain customers.
Notion’s published timeline says Fireworks told it on November 2 that a small set of data had been accessed by malicious actors; Notion initially understood the material to be synthetic testing data. On November 3, Notion says it disconnected from Fireworks, deleted data in the Fireworks system and halted training runs. Fireworks later told Notion that customer data might also have been accessed. Notion updated its advisory on November 10 and November 11 to describe the chat transcripts and key revocations. BleepingComputer reported the expanded disclosure on November 11.
Notion says it is auditing critical vendors, including inference providers and data processors, to check data-minimization practices and zero-retention policies. Its notice does not provide technical details about the system that retained the data.
“Unfortunately, during our investigation we learned that data was retained from the Fireworks endpoints that we used.”
— Notion, in its updated Help Center notice
Key Questions About Retention Remain
Notion has not said which Fireworks system retained the data, how long it was stored, or how many people and records were involved. The published information also does not explain why data covered by the stated contractual zero-retention commitment was retained, or whether it remained accessible after Notion disconnected from Fireworks.
The relationship between Notion’s account and Fireworks’ bulletin is not fully clear. Fireworks said it had not detected an impact to several services described as zero-retention by default, while also reporting access to datasets for certain customers. The companies have not publicly identified the specific endpoint or system involved in the Notion data exposure. It is also unknown whether other Fireworks customers experienced a similar retention issue.
Notion’s estimate covers workspaces, not a disclosed count of users or transcript entries. Its statement that most affected transcripts were unattributable describes missing labels; it does not confirm that all text was harmless or impossible to identify from its contents.
Vendor Audit and Customer Reviews
Notion says it is auditing critical vendors for compliance with data-minimization and zero-retention practices. The company has not announced a date for completing that review or said whether it will publish findings about Fireworks’ system and retention period. Further detail from Notion or Fireworks would help clarify how the exposed dataset was stored and what safeguards failed.
For customers, the next step depends on whether Notion has identified their workspace as affected. Notion says a workspace that has not been contacted has not been identified as affected so far. Affected organizations should follow the company’s security guidance, review potentially exposed chat content and rotate credentials that were pasted into Notion AI. The scope and cause remain subject to further disclosure.
Key Questions
What Notion data was exposed?
Notion says a dataset accessed in the Fireworks breach included Notion AI chat transcripts used for Insights. It says transcripts included text written as input to or output from the model, and could contain sensitive information.
How many Notion workspaces were affected?
Notion estimates that fewer than 1% of workspaces were affected. It has not disclosed a count of individual users, messages or transcript records.
Were Notion AI Meeting Notes affected?
Notion said AI Meeting Notes, including audio, transcripts and summaries, were not affected unless that material was used as input in a Notion AI chat. It said no other Notion features were affected.
What should an affected workspace do?
Notion advises affected customers to treat the relevant data as exposed, review it using the company’s Security incident review guidance, and rotate passwords, API keys or other credentials pasted into Notion AI during the incident. Notion said it revoked exposed Notion API keys found in affected messages on November 11.
What is still unknown about the breach?
Notion has not identified the Fireworks system that retained the data, the retention period, the number of affected people or records, or whether other Fireworks customers faced a similar issue.
Source: ThorstenMeyerAI.com
