Sixty-Four Million Job Applicants' Files Were Protected by a Password No One Had Bothered to Change
What happened
Researchers discovered in June 2025 that the administration interface for McHire.com, McDonald's AI-powered recruitment platform built on Paradox.ai's Olivia chatbot, was accessible through credentials so weak they offered no real protection. The platform had been processing job applications at global scale for an extended period before the flaw was reported.
The exposed data covered up to 64 million applicants. Names, home addresses, phone numbers, email addresses, the specific role each person had applied for, resume details, full work histories, personality test results, and transcripts of their AI-conducted hiring interviews were all reachable without authorization. The scope of what was exposed reflects what modern AI hiring tools are built to do: not just collect contact information, but gather behavioral and psychological data at each step of the application process.
The failure had two components. Technologically, the administration interface relied on weak default credentials, and the platform's APIs lacked sufficient access controls. Operationally, neither the vendor nor McDonald's had verified, after deployment, that security measures matched the sensitivity of the data flowing through the system. McHire.com is not an experimental product. It is the primary hiring pipeline for one of the world's largest hospitality operators, processing millions of applications across many countries.
After the breach was disclosed and patched, McDonald's attributed the failure to Paradox.ai, which acknowledged the oversight. That attribution deserves scrutiny. Job seekers who submitted resumes, completed personality assessments, and had their interview responses recorded did so under McDonald's branding. When a company deploys a third-party AI platform to conduct the first stages of hiring, it takes on a responsibility for how that platform protects applicant data. The fact that a vendor built the tool does not transfer the obligation to the vendor alone.
The breach was patched quickly, and no malicious exploitation has been confirmed. But the exposure window is unknown, and the only way the vulnerability came to light was through external research, not internal monitoring. Nothing in the record suggests any audit or continuous check flagged the misconfigured credentials before they became news. That is the gap this incident makes visible: a provable record of what a deployed system could access, who had last verified its configuration, and under what conditions data was being served would have caught a default password before a researcher did, and would have established, after the fact, exactly when the window opened and who bore responsibility for it.
Reported impact
- Affected parties
- Not publicly disclosed
- Harm type
- Not publicly disclosed
- Scale
- Not publicly disclosed
- Financial impact
- Not publicly disclosed
- Regulatory action
- Not publicly disclosed
Classification
Relevant governance controls
Governance control mapping is not available for this record.
- No controls mapped
Not publicly disclosed
Control mapping is analytical. It does not state that any control would have prevented the incident.
Sources and evidence
This record was researched and written by the Index. The event is also catalogued in the following database, which is listed for cross-reference.