Skip to content

Leak of Uber drivers' licences: Fireworks AI breach opens a gap in model supply chains

Share
Leak of Uber drivers' licences: Fireworks AI breach opens a gap in model supply chains

Listen to this article

Read by Anchor

The leak of driving licences for thousands of Uber drivers did not come this time from a direct breach of its central systems, but through a back-door in AI supply chains. The company confirmed that sensitive data concerning its drivers was stolen after a security breach at the Fireworks AI platform, a technology provider that enables developers to run open-source language models, after attackers seized internal credentials that gave them unauthorized access to the company’s records stored with the external provider.

The legal notices Uber sent this week to prosecutors in several U.S. states, including California, Illinois, Massachusetts, South Carolina, and Rhode Island, documented the initial scale of the damage. The stolen data included driving licences for 2,567 drivers in Massachusetts, 3,430 individuals in South Carolina, and 1,873 residents in Rhode Island, with references to affected parties in Illinois and California whose final numbers have not yet been determined, amid assurances that the total count is likely to rise as more official notices are filed.

The financial motive behind the attack reflects a shift in cybercrime toward targeting language-model API keys and sensitive datasets to turn them into quickly tradable assets.Fireworks AI acknowledged that the incident, which dates back to last October, affected customer billing data and proprietary training datasets, confirming that the attackers aimed to sell computational access keys, while the platform declined to discuss the vulnerability in detail and has not yet publicly identified any cyber-crime group as responsible.

This breach puts Uber’s security record under test for the second time this year, after the Uber Freight trucking unit suffered a prior ransomware attack by an organized cyber gang. The picture is reinforced by Notion’s admission that its data were also stolen in the same attack, underscoring that targeting cloud intermediaries and model-hosting providers has become the fastest route to penetrate the deep infrastructure of the world’s largest platforms with a single focused strike.

The development places technology leaders and engineering teams in the Gulf and Egypt under urgent security and regulatory pressure, especially in ride-sharing, logistics and fintech sectors that rely on language-model intermediaries to automate document verification and data extraction. Transmitting national ID documents and driving licences through cloud inference APIs without prior sanitisation shifts the entire breach risk to the company and subjects it to the strict penalties outlined in personal data protection laws in Saudi Arabia and the United Arab Emirates, prompting the isolation of identity data in local processing environments before any programmatic queries are shared with external AI providers.

Don't miss the next story

Subscribe for updates