AI in Hiring Notice

Inleika uses artificial intelligence to help people find work and help organizations hire. This notice explains:

• what our AI does and does not do;
• what data it uses;
• its limitations;
• how people stay in control.

Part A is for everyone. Part B is additional information for organizations.

Part A: For candidates and everyone else

1. Our principle: AI assists, people decide

AI in Inleika produces suggestions, rankings, explanations and summaries. People at the hiring organization make every decision: whether to contact, interview, offer or hire. Inleika never decides for them, and the AI never rejects anyone automatically.

2. Where AI is used

3. What data the AI uses

• Professional profile: title, work history, education, skills, languages, years of experience, summary.
• Preferences: job type, work arrangement, availability, location and salary expectation, used to check compatibility with the position.
• Position: the position's description, requirements and conditions, as written by the organization.
• Interview evidence: evaluator notes and Mila transcripts, for interview summaries only.

What the AI is not given or asked to infer:

• protected characteristics such as age, sex, ethnicity, religion, disability or marital status;
• your photo;
• facial or voice characteristics.

No profiling of emotions or personality.

4. How ranking works, and what it means for you

When an organization runs matching or a search, the system:

1. finds candidates whose skills relate to the position's requirements, using the skills vocabulary;
2. scores the strength of each relationship;
3. uses AI to assess and explain fit.

The most relevant candidates are shown first, and only a limited number are shown for each run. If you rank lower, an organization may not see you for that position. Ranking depends on how your profile describes your experience, so keeping it complete and accurate helps.

5. Limitations

• AI can misread a CV, miss transferable experience, or overstate a match. That is why you approve your profile details, and why organizations must review the AI's output before acting.
• Explanations describe the factors that drove a match. They are not a guarantee of suitability.
• Skills that are not in the vocabulary yet may be matched less precisely.
• AI interviews end without a conclusion when there isn't enough evidence. Insufficient evidence never counts as a negative result.

6. Fairness

• Our prompts and rules exclude protected characteristics and forbid final hiring decisions.
• Every AI result is stored with its reasoning, so it can be reviewed later.
• We test our matching for unequal outcomes and fix problems we find.
• Organizations must not use our AI to discriminate (see the Business Terms).

7. Your choices and rights

You can:

• turn discovery off so your profile is not included in marketplace matching;
• correct your profile at any time;
• ask how matching works for you, and object to automated processing;
• ask a person to review a result that affected you. Write to privacy@parzr.com. For a decision an organization made, we will pass your request to it;
• decline or stop an AI interview.

Section 9 of the Privacy Notice explains response times and complaints.

Part B: For organizations

This part describes the Service's automated decision-making tools for organizations that use them as deployers.

8. Intended uses

• Surfacing and ranking candidates for a specific, real position.
• Explaining fit to help a qualified person decide whom to contact or interview.
• Preparing structured, job-related interviews and summarizing evidence for human evaluators.

9. Uses the tools must not be put to

• Making a hiring decision, or rejecting someone, without meaningful human review.
• Decisions about credit, insurance, housing, tenancy or background checks.
• Evaluating anything other than job-related qualifications, including protected characteristics, or proxies for them such as ZIP code.
• Monitoring employees or evaluating current workers' performance.

10. Data and models

• Models: Inleika does not train its own machine-learning models on personal data. It uses general-purpose AI models from third-party providers, through API calls, without training them on your data.
• Skills vocabulary: built from an open-source technology skills catalog and a curated set of soft skills, with relationships reviewed for consistency.
• Inputs: the candidate data listed in section 3 and the position data your organization provides.

11. Meaningful human review

• Read the reasons and "points to verify", not only the score.
• Look beyond the top results when a position has few strong matches.
• Use the suggested interview questions to verify claims rather than to confirm the score.
• Record your own assessment before acting.
• Give a candidate a human review when they ask for one.

12. Your notice duties

Depending on where you hire, you may have to tell candidates that AI is used, before or after decisions. Examples:

• notice under the Illinois Human Rights Act;
• notice and a published bias audit under New York City Local Law 144;
• notices and adverse-outcome explanations under Colorado's automated decision-making law, from January 1, 2027.

Inleika provides template notices you can adapt. Contact privacy@parzr.com for them. You remain responsible for your notices and decisions.

13. Changes and records

We will tell organizations about material changes to these tools before they take effect. We keep previous versions of this notice for at least 3 years.