AI Disclaimer

How AI Analysis Works

VividCV uses Anthropic's Claude AI to compare candidate CVs against a provided job specification. The AI extracts relevant information from both documents and scores candidates across four dimensions: skills match, experience level, industry alignment, and seniority. These scores are combined into an overall match score out of 100.

The AI also generates a written summary for each candidate explaining the key reasons behind their score. This summary is intended to save time — not to be the final word on a candidate's suitability.

Limitations of AI Scoring

• AI can make mistakes. The model may misread, misinterpret, or overlook information in a CV or job spec. Always review the source documents alongside AI scores.
• Scores reflect what is written, not who the person is. A well-written CV may score higher than a poorly presented one even if the underlying candidate is stronger. CV quality is not the same as candidate quality.
• AI does not know your full context. The job spec you upload may not capture every requirement. The AI scores against what it is given.
• Results vary with input quality. Vague job specs or incomplete CVs will produce less reliable scores. The more specific the input, the more useful the output.
• AI is not infallible on niche roles. For highly specialised or unusual roles, AI scoring may be less reliable than for common role types.

Bias & Fairness

AI models can reflect and amplify biases present in training data. While Anthropic works to reduce bias in Claude, no AI system is fully free of it. VividCV scores are based on the textual content of CVs — however, indirect signals in that content (such as career gaps, certain qualifications, or writing style) may correlate with protected characteristics.

You must not use VividCV scores as a basis to discriminate against candidates on protected grounds including age, disability, gender, race, religion, national origin, marital status, sexual orientation, or any other characteristic protected by applicable law. AI scores are one input among many — not a justification for discriminatory decisions.

Your Responsibilities

• Always apply human review before shortlisting, rejecting, or advancing any candidate.
• Treat AI scores as one data point — consider the full CV, interview performance, references, and your own judgement.
• Ensure your use of AI-assisted screening complies with employment law in your jurisdiction.
• Inform candidates if AI tools are used in the screening process where required by law.
• Do not share AI summaries externally as official assessments without your own review and validation.

Questions

If you have questions about how AI is used in VividCV or want to raise a concern, contact us at support@vividcv.ai.