How Our AI Works — Transparency Notice

This notice explains what HRGuru's AI does, what it does not do, the safeguards built in, and your rights — as required by the EU AI Act and GDPR. AI assists recruiters; it does not replace human decision-making.

1. Classification

HRGuru's CV-scoring system is a HIGH-RISK AI system under the EU AI Act (Regulation (EU) 2024/1689), Annex III, point 4(a) (recruitment or selection of natural persons — analysing and filtering job applications and evaluating candidates). We maintain technical documentation, logging, and human-oversight measures appropriate to that classification.

2. What the AI does

• Parses CV text and extracts experience, skills and qualifications.
• Compares the candidate profile against the role's required skills, level and criteria, as defined by the employer.
• Produces a 0–100 score, a seniority classification, a recommendation, identified strengths and gaps, and a plain-language explanation.

2.1 Criteria the AI uses

The model evaluates only job-relevant factors: skills and technologies relevant to the role; depth and relevance of work experience; seniority indicators (scope, responsibility, achievements); education and qualifications where relevant.

3. Built-in human-oversight safeguard

The "Controversial" flag. HRGuru scores each CV through several independent AI evaluator perspectives. When the spread between those evaluators exceeds 30 points, the candidate is automatically flagged "Controversial", which signals that the case is ambiguous and requires mandatory human review before any action. This safeguard keeps a human in the loop precisely where the AI is least certain.

4. What the AI does NOT do

• It does not make the hiring decision — a human recruiter always decides.
• It does not use protected characteristics (age, gender, ethnicity, religion, disability) as criteria.
• The AI score is one input among several and is never the sole basis of a decision.
• It does not retain a CV beyond the configured retention period (default 365 days).
• It does not learn from or store candidate data for third-party model training (API processing only; training disabled where the option exists).

5. Human oversight

Every AI score is advisory. The employer must keep a human in the loop; recruiters can override the score, and the Controversial flag forces human review for ambiguous cases. HRGuru is designed so that AI scores are advisory and subject to human review. Where a score would nonetheless be relied on decisively (see CJEU C-634/21), candidates have the rights under GDPR Art. 22: to obtain human intervention, to express their point of view, and to contest the assessment — via the hiring company or privacy@hrguru.work.

6. Your rights

• Right to be informed that an AI tool is used (this notice; EU AI Act Art. 13, GDPR Art. 22).
• Right to human review of an AI-assisted decision — request this from the hiring company.
• Right to an explanation of a decision that affected you — from the hiring company or via ai@hrguru.work.
• Right to object to automated processing (EU/EEA, GDPR Art. 21).

Candidates exercise these rights with the employer (the controller); HRGuru assists the employer as processor. Contact: ai@hrguru.work · privacy@hrguru.work.

7. Bias testing

We monitor scoring for bias and maintain technical documentation. We intend to publish independent bias-audit results, and aim to complete the first independent audit by [target date — e.g. Q4 2026]. Employers using HRGuru as an AEDT for NYC candidates must conduct their own annual bias audit under NYC Local Law 144; HRGuru provides reasonable supporting information (see the DPA).

Legal review required: set the bias-audit target date and confirm the AI Act references with counsel before publication.