🔍 Read the full analysis: Notion Says Fireworks Breach Exposed AI Chat Transcripts, Despite A Zero-Retention Promise on ThorstenMeyerAI.com
Get business pricing on tech for your team
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
Notion says a dataset accessed in the Fireworks AI breach included stored Notion AI chat transcripts used for its Insights feature. The company says fewer than 1% of workspaces were affected and that it revoked exposed Notion API keys found in affected messages on November 11. Notion says Fireworks had promised zero data retention, but data was retained; the system involved and retention period have not been disclosed.
Notion says a dataset accessed in the Fireworks AI breach included stored chat transcripts from its AI service, and that the data had been retained by Fireworks despite a contractual zero-data-retention commitment. Notion estimates that fewer than 1% of its workspaces were affected and says it revoked exposed Notion API keys found in affected user messages on November 11.
The transcripts were associated with Notion’s Insights feature, which gives workspace administrators information about how teams use AI. Notion says it sent stored requests and responses from Notion AI use to Fireworks for analysis. A dataset Fireworks provided to Notion on November 8 included chat transcripts from that work, according to Notion’s updated security advisory.
Notion says most affected transcripts were unattributable: they were not labeled with the author or workspace and did not include tool-call results. However, the company defines a transcript as any input to or output from the model written to the chat. That means information users entered in a chat could be present, and Notion advises affected customers to review the data for sensitive content.
Notion says fewer than 1% of workspaces were affected. It advises those customers to treat the 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. The company also says it revoked exposed Notion API keys found in affected user messages on November 11.
Notion AI transcripts exposed after Fireworks breach
Notion says a dataset accessed in the Fireworks AI breach included stored chat transcripts used for Insights, despite a zero-retention commitment. Affected customers should review exposed content and rotate credentials shared in chats.
What was exposed
The dataset included stored requests and responses from Notion AI use for the Insights feature.
Chat transcripts
Notion defines a transcript as model input or output written to chat, so user-entered information could be present.
Often unlabeled
Most reviewed transcripts were not labeled with an author or workspace. That does not show the content was harmless.
What remains unclear
The public notices leave key questions about the retention incident unanswered.
- Which Fireworks system retained the data
- How long it was held
- How many users or transcripts were involved
- Why data was kept despite the commitment
- Whether other Fireworks customers were affected
- How the retained dataset relates to Fireworks’ service statements
What customers should do
Notion advises affected workspaces to treat the data as exposed.
Follow Notion’s incident-review guidance and look for sensitive content.
Change passwords, API keys, or other credentials pasted into Notion AI.
Notion is auditing critical vendors; the technical cause and retention period remain unpublished.
What the Chat Exposure Means
The incident matters because chat transcripts can contain more than routine prompts: users may enter customer names, contract details or credentials. Notion says the affected data should be treated as exposed, while also noting that most transcripts it reviewed were not labeled with a person or workspace. That lack of labels may limit direct attribution, but it does not establish that the content itself was harmless.
The disclosure also raises questions for organizations that choose AI services based on data-retention assurances. Notion says Fireworks made a contractual zero-retention commitment for its use of the platform, yet data from the endpoints Notion used was retained. The company has not said where or how that occurred, leaving customers without a public explanation of the gap between the commitment and the incident.
For affected workspaces, the immediate concern is identifying sensitive information in the exposed material and changing any credentials that appeared in chats. For other organizations, the case shows why vendor assurances and the data flows behind AI features may need to be reviewed alongside the product’s own privacy settings.
Products Worth Considering
Samsung 27-inch Odyssey OLED G6 QHD Gaming Monitor
LG UltraGear 34G600A-B 34-Inch Curved WQHD Gaming Monitor, 160Hz
Samsung Odyssey G9 QD-OLED 49 Dual QHD Curved Gaming Monitor
As an affiliate, we earn on qualifying purchases.
How Notion Used Fireworks
Notion says it used Fireworks for two purposes: classification jobs for Insights and general inference for open-weight models available within its product. The exposed data described in Notion’s notice came from the Insights work, which involved stored requests and responses rather than only a live request sent to generate an answer.
Notion’s advisory distinguishes that use from its zero-retention commitments for live requests to third-party model vendors. The company says those commitments apply to API requests used to generate Notion AI responses, but are not intended to cover data that must be retained for features such as chat history and Insights. Separately, Notion says Fireworks gave it a contractual zero-retention commitment for its platform use, then states that its investigation found data had been retained from the Fireworks endpoints it used.
Fireworks’ own update, last marked November 9, said it had not detected an impact to Inference, Training SDK and Nexus, which it described as zero-retention by default. The same update listed datasets among data accessed for certain customers. Notion has not identified the Fireworks system that held its data, and the published statements do not explain how the retained dataset relates to Fireworks’ description of affected services.
“Unfortunately, during our investigation we learned that data was retained from the Fireworks endpoints that we used.”
— Notion
What the Notices Do Not Explain
Neither company has publicly identified which Fireworks system retained the data, how long it was held, or how many individual users’ messages were involved. Notion gives an affected-workspace estimate—fewer than 1%—but has not published a number of people or transcripts.
The public notices also do not explain why data was retained despite the commitment Notion says Fireworks made, or whether the retention resulted from a particular system, configuration or process. Fireworks’ statement about zero-retention services and its account of accessed datasets do not resolve that question. It is also unclear whether the same issue affected other Fireworks customers.
Notion says most affected transcripts could not be tied to a person or workspace, but that does not answer whether sensitive details in the content could identify users or organizations. The company has urged affected workspaces to examine the material rather than assume it contained no sensitive information.
Vendor Review and Customer Checks
Notion says it is auditing critical vendors, including inference providers and data processors, to check their data-minimization practices and adherence to zero-retention policies. It has not announced a completion date for that review or published further technical details about the Fireworks data store.
Customers who have received notice should follow Notion’s incident-review guidance, assess the affected data and rotate credentials that may have been entered into Notion AI. Notion says it has not identified workspaces as affected if neither the workspace owner nor the user has received a notice, but customers can consult the company’s updated advisory for current instructions.
The next material updates would be a fuller account from Notion or Fireworks identifying the system involved, the retention period, and whether other customers were affected. Until those details are published, the scope and technical cause of the retention remain unresolved.
Key Questions
What Notion AI data does Notion say was exposed?
Notion says the dataset accessed in the Fireworks breach included chat transcripts used for its Insights feature. It defines a transcript as model input or output written to the chat, so content users entered may have been included.
How many Notion workspaces were affected?
Notion estimates that fewer than 1% of workspaces were affected. It has not published the number of individual users or transcripts involved.
Does Notion say every transcript identified its author?
No. Notion says transcripts for the vast majority of affected workspaces were unattributable, meaning they were not labeled with the person who wrote them or the workspace they came from. The company still recommends reviewing the content for sensitive information.
What should an affected Notion workspace do?
Notion advises affected customers to treat the data as exposed, review it using its Security incident review guidance, and rotate passwords, API keys and other credentials that may have been pasted into Notion AI during the incident.
What remains unknown about Fireworks’ retention?
The companies have not publicly said which Fireworks system retained the data, how long it was stored, why retention occurred despite the contractual commitment, or whether other Fireworks customers were affected.
Source: ThorstenMeyerAI.com
