Uber driver licence data stolen after hackers breach AI supplier Fireworks AI
Uber has confirmed that almost 150,000 people had personal information compromised after attackers broke into one of its AI suppliers. A filing with Oregon's attorney general puts the number of victims at 149,346. The stolen data included images of drivers' licences and details taken from them: names, home addresses, dates of birth and licence numbers.
Uber does not name the supplier in its filings, but told TechCrunch it was Fireworks AI, a company that provides infrastructure for developers to run and fine-tune AI models. According to a notification filed with Iowa's attorney general, Fireworks alerted Uber to the intrusion on October 9. Uber cut off the supplier's access and, four days later, established that attackers had accessed drivers' personal information on October 7.
Uber had set up a dedicated cloud storage environment for the supplier, with security monitoring run by its own cloud security team. However, the attackers were using legitimate credentials, so no alarms were triggered. Uber says those credentials could not reach its other systems, and that it has since revoked or changed the credentials tied to the supplier. It has not explained why it gave licence images to an AI inference company.
Fireworks had a difficult October. It first disclosed an incident on October 9 after finding someone using its internal credentials without permission, and said it had no evidence personal information or customer datasets were taken. After a second compromise later that month, it disclosed on November 9 that billing records, some customer datasets and incomplete model weights had been accessed. Fireworks suspects financially motivated criminals who sell unauthorised access to large language models.
Uber began notifying people on November 10, more than a month after learning of the breach. It is offering free credit monitoring and fraud assistance, but drivers have only 90 days from the date of their letter to sign up.
Why it matters
Uber's monitoring was in place and still failed, because activity using valid supplier credentials looked normal. This breaks the assumption that a segregated, monitored environment makes sharing sensitive data with a third party safe. Any business passing identity documents to AI vendors has to decide whether the data needs to leave its control at all, and how to detect abuse of legitimate logins.
What to do
- Review which suppliers, especially AI vendors, hold identity documents or personal data and remove anything they do not need.
- Revoke or rotate supplier credentials promptly after any supplier reports an incident.
- Confirm supplier credentials are scoped so they cannot reach your other systems.
- Add alerts for unusual access patterns by supplier accounts, since valid logins will not trip standard intrusion alarms.