Fireworks Breach Reaches Uber: 150,000 Drivers Exposed

Human Verified 0

Share to: , and

Dark rainy city street with a lone car beside a glowing data center, red warning lights marking a security breach

Uber is notifying nearly 150,000 drivers that their personal information was stolen after attackers reached it through Fireworks AI, an artificial-intelligence vendor the ride-hailing company used. The exposed data includes names, driver’s license numbers, dates of birth and home addresses.

What Uber Told Drivers

Uber sent letters to drivers on November 10. The letters said attackers accessed personal information through an Uber account held with a third-party service provider. Uber did not name the provider in the letters, and it gave no total number of affected drivers there.

The scale emerged in a filing with Maine’s Office of the Attorney General, which puts the number of exposed drivers at 149,346. It lists the same four fields for each: full names, driver’s license numbers, dates of birth and home addresses.

How the Attackers Got In

Uber’s timeline starts on October 9, when the provider told it attackers had compromised its internal credentials and used them to reach Uber’s account. Uber said it cut its data connections with the provider at once and opened an investigation.

On October 13, the investigation confirmed the attackers had used the stolen credentials to access Uber data held with the provider. Uber has not explained why driver identity documents were reachable through an artificial-intelligence vendor’s systems, and that question remains open.

Fireworks detected unauthorized use of its internal credentials on October 6, and blames a financially motivated group linked to large-language-model hijacking. A second intrusion followed on October 23 and 24, when a cloud provider hosting Fireworks research machines was itself breached.

Uber said it had alerts switched on to catch unauthorized access to its Fireworks environment. None fired, because the attackers moved with Fireworks’ credentials. That shows why stolen vendor keys are hard to spot: to every alarm, the intruder looks like the vendor itself.

It is not yet clear which of the two intrusions exposed the driver records. Fireworks says the second incident reached development machines and exposed billing information, secrets and incomplete model weights for some customers. It is working with Palo Alto Networks’ Unit 42.

A Wider Pattern

Uber is the third company known to have lost data in the Fireworks breach. Notion said its artificial-intelligence chats leaked, and Cursor traced its own breach to a leaked Fireworks key, with attackers infecting some users’ devices with backdoor malware.

Fireworks sells developers fast access to open artificial-intelligence models and the infrastructure to run them. It says more than 10,000 companies use its platform today, with Uber among its customers. That reach turned one vendor’s breach into several companies’ problem.

Our Read

This is the supply-chain breach in its purest form. Uber’s systems were not the entry point; the keys to its data sat with a vendor, and the vendor’s keys were stolen first. Companies audit their own defenses, but data held by suppliers rarely gets the same scrutiny.

The unanswered question matters more than the count. Driver license numbers and birth dates sat in a vendor’s environment, and nobody has said why. Until vendors explain what customer data they hold, every new AI supplier is a new place for sensitive data to leak.

Why It Matters

License numbers, birth dates and home addresses are the raw material of identity theft, and they cannot be changed like a password. For 149,346 drivers, this breach is not a one-week problem. It also shows an AI vendor’s failure reaching far beyond the AI industry.

Background

Uber has been breached before. Hackers who broke into its systems in October 2016 stole data from 50 million users and 7 million drivers, and the Lapsus$ group hacked it in September 2022. This is its second breach this year, after its Uber Freight unit was hit.

Source: TechCrunch’s report and BleepingComputer’s report

Share to: , and

Faheem R. Khattak

Senior automation engineer and AI marketing journo covering LLMs, algorithms, analytics, core updates, Bing search, and technical SEO.

0 Comments

Leave a Reply

Optional Take Tag:

🛡️ Spam protection: Honeypot + Time-Trap Active

This site uses Akismet to reduce spam. Learn how your comment data is processed.