kvyk.ai

Content moderation policy

Last updated:

kvyk.ai lets an organisation build its own business apps with the help of an AI assistant, and run them for its own people and customers. This policy says what the assistant and the apps may and may not be used for, how we watch for misuse, and what we do when we find it. It is part of the acceptable use rules in our terms and conditions.

1. What kvyk.ai is for

  • Building and running business apps: keeping records, forms, overviews, documents, automations and messages for the organisation’s own work.
  • Asking the assistant to design, change and explain those apps, and to answer questions about the organisation’s own data, with the access of the person asking.
  • Letting an Ask the AI step in an automation read and decide, such as sorting a message into a category or summarising a text.

2. What is not allowed

The rules of section 5 of the terms apply to everything on kvyk.ai, the assistant and the apps included. Nobody uses kvyk.ai:

  • for anything illegal, or to make, store or share content that is illegal or infringes the rights of others;
  • to phish or deceive: apps or messages that pretend to be someone else, or that collect passwords, payment details or personal data under false pretences;
  • to spread malware, or to send spam or other unsolicited messages;
  • to make or spread content that sexualises minors, that threatens or harasses people, or that incites violence or hatred;
  • to try to break, overload or get around the security of the service, to reach the data of other organisations, or to get the assistant to ignore its rules;
  • to resell the service without our written agreement.

3. Sensitive and high-risk uses

The assistant is a tool for building apps. It is not an adviser, and its output can be wrong.

  • Do not rely on the assistant, or on an Ask the AI step, for medical, legal or financial advice.
  • Do not let AI alone make a decision that has legal or similarly significant effects on a person, such as hiring, dismissal, credit, insurance or access to essential services. A person decides.
  • Do not use kvyk.ai for practices the EU AI Act prohibits, such as social scoring, or recognising the emotions of staff or pupils.
  • kvyk.ai is not built for special categories of personal data, such as health data, religion or political views. As the data processing agreement says, an organisation stores such data only after it has judged that the measures there are enough for it, and only with a lawful ground to process it.
  • Apps that reach the public make clear who is behind them, and say so when people deal with AI rather than with a person.

4. How we watch for misuse

  • Anyone can report an app at abuse@kvyk.eu; see Report abuse. We read and answer every report.
  • Automatic safeguards run all the time: every file added to kvyk.ai is scanned for malware before anyone can open it, signing in and repeated requests from one source are limited, and each organisation’s use of the assistant and of Ask the AI steps is measured against its plan.
  • The assistant acts only as the person chatting, with that person’s access, and asks for their confirmation before anything destructive, outward or on Live.
  • We do not read customer data as a routine. When a report, a safeguard or the law gives us reason, the people at kvyk.ai who are bound to confidentiality may look into the organisation, its apps and the technical records needed to establish what happened, within the data processing agreement.
  • We work with the authorities where the law requires it.

5. What we do when the rules are broken

Depending on how serious it is, we may:

  • warn the organisation’s Administrators, or the person concerned, by email and ask them to put it right;
  • suspend a person: they are signed out at once and cannot sign in until the suspension is lifted;
  • suspend the organisation: nobody can work in it, its apps do not open and its customers cannot sign in, until the suspension is lifted;
  • end the agreement, as section 12 of the terms sets out: at once when a serious breach is not put right within 14 days of our asking;
  • report to the authorities where the law requires it.

We tell the organisation first and give it a chance to put things right, unless the breach is serious or we must act at once to protect the service, other customers or the people concerned. A suspension is never a deletion: nothing is deleted, and we email the reason. Each warning and suspension is recorded in the organisation’s event log. We lift a suspension once the cause is gone.

6. If you disagree

An Administrator or a person who thinks we acted wrongly writes to privacy@kvyk.eu, the address in our email. We look at the decision again, with what they tell us, and answer.

7. Changes

We change this policy as the service and the law change. The date at the top says when it was last changed. Because the policy is part of the terms, a change to it is a change to the terms, made as section 13 of the terms sets out.