Mercor

Company overview

Mercor (Mercor.io Corporation) is a San Francisco-based AI hiring company founded in 2023 by Brendan Foody, Adarsh Hiremath, and Surya Midha, and is heavily venture-backed (a reported $350M Series C at a ~$10B valuation in October 2025). It operates as a two-sided platform: a contractor/expert marketplace where Mercor itself administers AI video interviews to vet specialists (doctors, lawyers, engineers, scientists) and place them with AI labs such as OpenAI and Anthropic, and an employer-facing enterprise product ('Top-of-Funnel Applicant Screening') that ingests ATS data and job descriptions to evaluate applicant pools against role-specific rubrics, producing calibrated scorecards and pass/fail recommendations. Both modes use a large language model to analyze a recorded video interview and generate a profile and predicted performance.

Regulatory exposure

Mercor's core function — automated evaluation, scoring, and ranking of people for employment and work placement — places it squarely in the highest-exposure category. Its AI video interview and applicant-scoring tools are high-risk under EU AI Act Annex III (employment), and the employer-facing screening product is an Automated Employment Decision Tool under NYC Local Law 144. The use of AI to analyze video interviews triggers the Illinois AI Video Interview Act, and capture of facial geometry and voiceprints implicates Illinois BIPA. As of mid-2026 Mercor publishes none of the artifacts these regimes expect: no public bias audit, no model/system card, no FRIA or deployer guidance, and a GDPR-only privacy policy that does not disclose automated decision-making, profiling, or candidate appeal rights. This exposure is not theoretical: a March 2026 supply-chain breach exposed video interviews, facial/voice biometrics, SSNs, and passports of 40,000+ people, and multiple class actions allege BIPA, Illinois AIVIA, and FCRA violations and that interview data was used to train AI models despite Mercor's public claims to the contrary.

Path to a higher score

Mercor could raise its score quickly by publishing the evidence that competitors already disclose: an independent NYC LL 144 bias audit (with auditor, date, and downloadable results), a system/model card and explainability statement for its interview-scoring and rubric-scoring models, EU AI Act deployer guidance and an Art. 27 FRIA template, and a data-governance page specifying training-data exclusions (including an explicit statement, with controls, that candidate interviews are not used for model training). Concrete human-oversight documentation — override controls, audit logs, and per-jurisdiction toggles — plus a public security/status page and incident channel would address both the regulatory gaps and the trust damage from the 2026 breach and litigation.