How We Use AI
Last updated: 8/19/2026
References are used to make decisions about real people, so we are deliberate about where artificial intelligence fits into our platform and where it does not. This page describes exactly what AI does at AxionCred, what it will never do, and how that maps to the laws governing AI in hiring.
What AI does on our platform
One thing: transcription. When a referee records a video reference, we use an automated speech-to-text service (OpenAI Whisper) to produce a written transcript. The transcript makes references searchable and accessible, and its length contributes to a “substantiveness” signal in the CRED score. That is the full extent of AI in the reference process.
What AI does not do
We do not use AI to score, rank, or evaluate candidates. We do not analyze faces, voices, emotions, or any other biometric characteristics in reference videos. We do not generate AI summaries that stand in for a referee’s actual words, and we do not use AI to infer personality traits, sentiment, or “fit.” Employers always watch the original video and read the full transcript — nothing is condensed, filtered, or interpreted by a model before a human sees it.
The CRED score is a formula, not a model
A CRED score is computed by a fixed, deterministic formula from five transparent dimensions: referee quality (seniority and relationship), reference depth, verification strength, portfolio breadth, and profile completeness. The same inputs always produce the same score. No machine learning is involved, the weights are inspectable, and no protected characteristic — or proxy for one — is an input.
Humans make the decisions
AxionCred verifies and organizes references; it does not recommend hiring outcomes. Employers on our platform agree to use CRED scores and references as one input among many, never as the sole basis for an employment decision. Candidates control who can view their references and can dispute inaccurate information.
Your data and AI training
Reference videos and transcripts are processed by our transcription provider solely to provide the service. Data sent through the OpenAI API is not used to train OpenAI’s models. We never sell reference data or use it to train models of our own.
How this maps to AI hiring laws
Because we do not use AI to assess candidates, AxionCred is designed to stay clear of the obligations that attach to automated hiring tools: we perform no AI analysis of video interviews of the kind regulated by the Illinois Artificial Intelligence Video Interview Act, no automated employment decision-making of the kind covered by New York City Local Law 144, and no candidate profiling of the kind restricted under the GDPR and the EU AI Act. If that ever changes, we will say so here first — plainly, and before the feature ships.
Questions
Ask us anything about our use of AI at privacy@axioncred.com. See also our Privacy Policy and FCRA Rights pages.