1. Introduction
This AI Recruitment Fair Use Addendum applies specifically to Halo, Aivry's recruitment intelligence product.
This document supplements:
Where conflict exists, stricter obligations apply.
2. Purpose of Halo
Halo is designed to support recruitment workflows. Halo may assist with:
- Resume parsing
- Candidate organization
- AI-assisted screening support
- Candidate scoring
- Skills matching
- Workflow automation
- Recruitment process acceleration
- Data organization
Halo is intended as a business productivity tool. Halo is not a replacement for lawful human hiring judgment.
3. Nature of AI Processing
Halo may process:
- Resumes
- Candidate applications
- Employment history
- Education details
- Candidate metadata
- Skill information
- Communication history
- Ranking outputs
- AI-generated assessments
Processing may involve:
- Parsing
- Classification
- Scoring
- Relevance matching
- Recommendation logic
- Workflow automation
Outputs may be influenced by:
- Uploaded customer data
- Prompt design
- Configuration logic
- Third-party AI model behaviour
- Incomplete source information
4. No Final Hiring Authority
Halo does not make final employment decisions. Halo does not:
- Hire candidates
- Reject candidates autonomously
- Execute final employment actions
- Determine legal eligibility
- Provide employment law compliance guarantees
All hiring decisions remain the sole responsibility of the customer.
5. Human Oversight Requirement
Customers must maintain meaningful human oversight. Human review is required before:
- Rejection
- Shortlisting
- Interview progression
- Compensation decisions
- Offer issuance
- Role elimination
- Adverse employment decisions
AI output must not be treated as automatic truth.
6. Bias Risk Disclosure
AI systems can produce biased outcomes. Bias risk may arise from:
- Model training limitations
- Historical hiring patterns
- Flawed customer data
- Incomplete resumes
- Prompt design
- Customer scoring logic
- Third-party model behaviour
Halo cannot guarantee bias-free output.
7. Customer Responsibility for Fairness
Customers remain responsible for:
- Fairness review
- Equal opportunity compliance
- Labour law compliance
- Employment law compliance
- Anti-discrimination obligations
- Internal governance
This includes avoiding discrimination based on protected characteristics where applicable. Examples may include:
- Race
- Religion
- Disability
- Age
- Gender
- Marital status
- Nationality
- Legally protected employment categories
Compliance requirements vary by jurisdiction.
8. No Explainability Guarantee
AI scoring may not always be fully explainable in deterministic human terms. Because AI models may be probabilistic, customers must avoid representing AI scoring as guaranteed objective truth.
9. Data Legality
Customers represent that they have lawful rights to process candidate data, including:
- Resume data
- Applications
- Personal identifiers
- Communications
- Recruitment records
Customers are responsible for lawful notices, consent where required, privacy compliance, and retention obligations.
10. Accuracy Disclaimer
Halo outputs may be:
- Inaccurate
- Incomplete
- Stale
- Contextually wrong
- Biased
- Irrelevant
- Misleading
Customers must independently validate outputs.
11. Automated Rejection Restriction
Unless explicitly contracted and legally reviewed, Halo must not be used for fully autonomous rejection decisions. Human review is required.
12. High-Risk Hiring Decisions
Halo should not be solely relied upon for:
- Executive hiring
- Regulated roles
- Sensitive roles
- Legally protected employment decisions
- Visa-dependent employment
- Employment eligibility determinations
Human review is mandatory.
13. Regulatory Responsibility
Customers remain solely responsible for complying with:
- Labour law
- Hiring law
- Anti-discrimination obligations
- Privacy law
- Employment fairness requirements
- Regulatory disclosure requirements
Aivry does not provide employment legal advice.
14. Prohibited Uses
Halo must not be used to:
- Unlawfully discriminate
- Auto-reject candidates without review
- Create exclusionary hiring pipelines
- Profile unlawfully
- Bypass fairness obligations
- Conceal automated decision logic where disclosure is required
15. Third-Party AI Dependency
Halo may rely on third-party AI systems. Capabilities may change due to:
- Model updates
- Provider changes
- API changes
- Outages
- Behavioural shifts
- Scoring variability
Static performance cannot be guaranteed.
16. Liability Allocation
Customers remain solely responsible for:
- Hiring outcomes
- Candidate treatment
- Compliance failures
- Discrimination claims
- Employment disputes
- Legal complaints
Aivry is not the employer.
17. Audit Recommendation
Customers are strongly encouraged to periodically review:
- Scoring logic
- Rejection patterns
- Fairness outcomes
- Process integrity
- Recruiter override behaviour
This is especially important for scale hiring.
