This policy forms part of the Terms of Service and applies to every employer account. Breaching it is a breach of the Terms.
We have written it as plainly as we can, because the obligations in it are the ones most likely to cause you real trouble if they are ignored.
1. Use it for hiring, lawfully
You may use JobsChat.ai to assess candidates for roles you are genuinely recruiting for.
You may not use it to:
- interview people for a role that does not exist, or that you have no intention of filling;
- gather information about a person for any purpose other than considering their application;
- assess your own employees without telling them what the assessment is for;
- build a database of people for sale, for marketing, or for any purpose other than your own recruitment;
- conduct research on people without their knowledge.
2. Do not discriminate
You may not use this platform to treat a candidate less favourably because of race, colour, nationality, ethnic or national origin, sex, gender identity, sexual orientation, pregnancy or maternity, marital status, age, religion or belief, disability, health status, or any other characteristic protected where you operate.
In particular, you may not:
- write a job post, a question, or an instruction to the platform designed to filter on, or draw out, any of the characteristics above;
- ask us to score or rank candidates on any such characteristic;
- use the interview recording to assess a candidate's appearance, accent, ethnicity, age or apparent health;
- use the platform to screen out candidates from a protected group by proxy — for example by using a characteristic that closely tracks one.
Our assessment is written from what a candidate said about their work. That does not protect you if you use the output improperly.
3. Offer an alternative to candidates who need one
This is an obligation, not a suggestion.
An AI video interview is not equally accessible to everyone. A candidate may have a speech impairment, a hearing impairment, a stammer, limited vision, a condition affecting concentration or processing, no access to a quiet space, no camera, or a poor connection. A candidate may not be a confident speaker of the interview's language.
Our speech recognition is imperfect and performs less well on some voices, accents and speech patterns than others. That is a limitation of the technology and we state it rather than hide it.
You must therefore:
- tell candidates how to request an alternative, in your invitation or in your careers process;
- provide a reasonable alternative — a human interview, a written exercise, more time, or a rescheduled attempt — where a candidate asks for one or where you become aware that they need one;
- not treat a request for an alternative as a mark against the candidate, and not record it as one.
If a candidate tells you the interview did not work for them, treat that as a reason to look again, not as a result.
4. A person must make the decision
You must have a person in your organisation review the assessment before any decision is taken about a candidate.
You may not:
- reject, advance, shortlist or rank out a candidate solely on our score, our recommendation, or the order of our list;
- configure any process of your own that acts automatically on our output;
- present our recommendation to a candidate as the reason for a decision without a person having considered it.
Our platform does not move candidates through your pipeline. Only a person clicking in your account can. Do not build around that.
5. Tell candidates what is happening
You must tell every candidate you invite:
- who you are, and that they are being considered by you;
- where your own privacy notice or candidate notice can be found;
- how to contact you about their data.
We disclose our part — the AI interviewer, the recording, the assessment, the retention, and who processes what. We cannot disclose yours. Your candidates have no relationship with us and no way of knowing who you are beyond the company name in the invitation.
6. Only upload data you are entitled to upload
You confirm, for every CV you upload, that you have a lawful basis for processing it and for sending it to a platform that will process it on your behalf.
Do not upload:
- a CV obtained without the person's knowledge, where your law requires it;
- anyone else's confidential or commercially sensitive information;
- content you have no right to share.
6.1 Sensitive information inside a CV
We recognise that in many markets a CV ordinarily carries a photograph, a date of birth, a nationality, and a marital status, and that an employer receiving one has no practical way to remove them before uploading. We are not asking you to do the impossible.
What we ask is this:
- Do not upload a CV that has been deliberately supplemented with sensitive information the candidate did not put there — a health report, a background check, a note about someone's religion or politics.
- Do not use any such information, however it reached us, as a basis for a decision. Section 2 already prohibits that.
- Do not ask the platform to read, infer, score or rank on it.
Our side of this: we do not extract a photograph, a date of birth, a nationality or a marital status into a candidate's structured profile. Where such details appear in the CV text itself, they are stored as part of that text and nothing is built on them.
Separately, special category data — health information, biometric data, trade union membership, religious or political beliefs, sexual orientation, or criminal records — must not be uploaded as a document or a field of its own unless you have a lawful basis for it and have told us first.
7. Interview links
An interview link is issued for one named candidate and works once.
Do not forward it to anyone else, publish it, or use it to interview a person other than the one it was issued for. The consent record attaches to the person the invitation names.
8. Do not attack or extract the platform
You may not:
- attempt to access any account, record or system that is not yours;
- probe, scan or test the security of the platform without our written permission;
- use automated means to extract data at scale;
- attempt to recover, reconstruct or reproduce our prompts, our question generation, our assessment methodology, or the models behind them;
- use output from the platform to train a competing system;
- resell or white-label the service without a written agreement with us.
If you find a security problem, tell us at security@jobschat.ai. We will not pursue anyone who reports one in good faith and does not exploit it.
9. Do not misrepresent what the assessment is
If you share an assessment inside your organisation, with a client, or with the candidate, do not present it as:
- a psychometric or clinical assessment;
- a measurement of intelligence, personality, integrity or character;
- a verified account of the candidate's history;
- a determination of suitability.
It is one AI-written reading of one conversation, quoting a transcript that may contain recognition errors.
10. What happens if you breach this policy
Depending on what happened, we may contact you, require you to stop, suspend your account, or close it. Where we are legally required to report something, we will.
Serious breaches — using the platform to discriminate, uploading data you have no right to, or acting automatically on our output — may end your account immediately.
11. Reporting
To report misuse of the platform, by anyone, write to support@jobschat.ai.
If you are a candidate and you believe an employer has used this platform improperly in relation to you, write to privacy@jobschat.ai. We will look at it.
JobsChat LLC, 2604 Whittier Pl, Wilmington, DE 19808, United States.