🔍 Read the full analysis: Fireworks Breach And Notion: A Closer Look At AI Transcript Exposure on ThorstenMeyerAI.com
Get hardware and tech essentials 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 data accessed in the Fireworks AI breach included Notion AI chat transcripts used for its Insights feature, affecting fewer than 1% of Notion workspaces. On November 11, Notion revoked exposed API keys it found in affected user messages; it has not said how long the data was retained or which Fireworks system held it.
Notion says a dataset accessed by attackers in the Fireworks AI breach included chat transcripts from Notion AI’s Insights feature, prompting the company to revoke exposed API keys found in affected messages on November 11. Notion says fewer than 1% of workspaces were affected, but the disclosure matters because the data was retained by Fireworks despite a contractual zero-data-retention commitment described by Notion.
Notion’s updated advisory says it received a dataset from Fireworks on November 8 that included Notion AI chat transcripts used for Insights, a feature that gives workspace administrators information about how teams use AI. The company says transcripts for most affected workspaces were “unattributable”: they were not labeled with the writer’s identity or workspace, and did not include tool-call results. Notion defines a transcript as any input to or output from the model that was written to the chat.
That distinction does not mean the text itself was harmless. Notion advises affected customers to review the data because chat inputs or outputs may contain sensitive information. The company says AI Meeting Notes were not affected, unless audio, transcripts or summaries were used as input in a Notion AI chat transcript. It also says no other Notion features were affected.
Notion says Fireworks had made a contractual commitment to zero data retention for Notion’s use of its platform. But during its investigation, Notion learned that data had been retained from the Fireworks endpoints it used. On November 11, Notion said it had revoked exposed Notion API keys found in affected user messages. It recommends that affected workspaces treat the data as exposed, review it through its Security incident review process and rotate passwords, API keys or other credentials included in chats.
Fireworks breach and Notion What transcript exposure means
Notion says data accessed in the Fireworks AI breach included Notion AI Insights chat transcripts. The incident highlights the gap between a zero-retention commitment and data retained for product features.
What data was involved
Notion says the exposed transcripts came from Insights classification work.
Chats could contain sensitive text
Inputs or outputs may include customer names, contract details or credentials.
Most were unattributable
Notion says most affected transcripts lacked writer and workspace labels and had no tool-call results.
Other features were not affected
AI Meeting Notes were excluded unless their content was used in an AI chat transcript.
What remains unclear
Published statements do not explain how the data was retained despite the contractual commitment.
What affected customers can do
Notion advises notified workspaces to treat relevant data as exposed.
Use Notion’s Security incident review process to examine affected chats.
Change passwords, API keys or other credentials included in chats.
Workspace owners should use Notion’s advisory for case-specific guidance.
Why Chat Transcript Exposure Matters
The disclosure makes a vendor-retention failure relevant to customers who may have used Notion AI for work conversations. A transcript can contain material users enter directly, including customer names, contract details or credentials. The absence of a name or workspace label may reduce direct identification, but it does not establish that the underlying content was non-sensitive.
It also highlights a practical limit of zero-retention assurances. Notion says its ZDR commitment for live requests to model vendors does not cover every feature that stores data for later use; it specifically identifies chat history and Insights as use cases that require retained data. For Insights, Notion says stored requests and responses were sent to Fireworks for analysis. The incident therefore raises questions about how data-retention commitments are applied across different product workflows, not just live model responses.
For affected users, the immediate concern is whether a message contained information that should be changed or protected. The exposure does not, on the information published so far, establish that every transcript was identifiable or that every exposed message contained sensitive material.
Products Worth Considering
Apple MacBook Pro 14.2-Inch Laptop with M6 Chip, Space Black
LG UltraGear 32-Inch QHD 180Hz Curved Gaming Monitor
Apple MacBook Air 13-inch Laptop with M5 Chip, 16GB Memory, 512GB SSD, Liquid Retina Display, and 12MP Center Stage Camera (2026, Silver)
As an affiliate, we earn on qualifying purchases.
How Notion and Fireworks Describe the Data
Notion used Fireworks for two purposes: classification jobs for Insights and as an inference provider for open-weight models in Notion’s product. Notion says the exposed transcript data came from the Insights work, not from its general inference use. The distinction matters because Fireworks’ own November 9 update says 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 information accessed for certain customers.
The public accounts do not identify which Fireworks system retained Notion’s data or explain how that retention occurred. Notion says it disconnected its systems from Fireworks on November 3, deleted data in the Fireworks system and halted training runs. Its updated advisory also reports that Fireworks first notified it on November 2 that a small amount of data had been accessed. These details describe Notion’s account of its response; Fireworks’ public bulletin, according to the source material, was last updated November 9.
Notion says it is auditing critical vendors, including inference providers and data processors, to check their data-minimization practices and compliance with retention policies. The breach has also been reported by BleepingComputer, which covered Notion’s disclosure on November 11.
“Unfortunately, during our investigation we learned that data was retained from the Fireworks endpoints that we used.”
— Notion
Unanswered Questions About Retention
The companies have not said which Fireworks system retained Notion’s data, how long it was stored, or why the retention occurred despite the contractual commitment. The available statements also do not establish the total number of affected people, as distinct from the share of Notion workspaces identified as affected.
It remains unclear whether other Fireworks customers experienced a similar retention problem. Fireworks’ statement about no detected impact to several zero-retention services does not explain where the dataset involving Notion was kept. The published information does not resolve that apparent gap, so it would be premature to infer that the same issue affected—or did not affect—other customers.
Notion says fewer than 1% of workspaces were affected. It has not provided a precise workspace count in the source material, and the statement does not specify how many transcripts or individual users were involved. Notion says users and workspace owners who have not been contacted have not been identified as affected, but the scope could change as the investigation proceeds.
Vendor Audit and Customer Reviews
Notion says its audit of critical vendors will examine data-minimization practices, including zero-retention policies. The company has not announced a completion date or published further technical findings about the Fireworks system involved. Further updates from Notion or Fireworks may clarify the retention period, the number of records affected and whether other customers need to take action.
Notion’s current guidance for customers who have been notified is to treat relevant data as exposed, examine it through Security incident review and rotate any credentials entered into Notion AI during the incident. Workspace owners should use Notion’s direct notices and advisory for case-specific instructions. The incident remains under investigation, and the published facts do not establish that every affected transcript contained sensitive information or that every credential in a message was misused.
Key Questions
What Notion data was included in the Fireworks breach dataset?
Notion says the dataset included Notion AI chat transcripts used for its Insights feature. The company says most affected transcripts were not labeled with an author or workspace, but the text could still contain sensitive information.How many Notion workspaces were affected?
Notion says fewer than 1% of workspaces were affected. It has not provided an exact count of workspaces, users or transcripts in the source material.What should a notified Notion workspace do?
Notion advises affected workspaces to treat the relevant data as exposed, review it through Security incident review, and rotate passwords, API keys or other credentials entered into Notion AI during the incident.Were Notion AI Meeting Notes affected?
Notion says AI Meeting Notes were not affected unless their audio, transcripts or summaries were used as input in a Notion AI chat transcript. It says no other Notion features were affected.Does the disclosure show that other Fireworks customers were affected?
No. The available statements do not establish whether other customers experienced the same issue. Fireworks said it had not detected an impact to certain services, but neither company has identified which system retained Notion’s data.Source: ThorstenMeyerAI.com
