Skip to content

Stockertim Dictionary · Artificial intelligence

Red line

Also called prohibited use, no-go use

In short

A red line is a use of AI that an organisation decides it will not allow, whatever the claimed benefit, because the harm is judged too great.

  • Beginner
  • Concept
  • Checked

In detail

A red line marks the edge of what an organisation is willing to do with AI. Most uses are judged case by case, balancing benefits against risks and safeguards. A red line is different: it is a firm rule that a particular use is off the table, decided in advance so that nobody has to argue it again under pressure.

Organisations draw red lines where harm to people, rights or trust is serious and hard to undo. Examples might include using AI to secretly monitor staff's private lives, to make final decisions about dismissals without human judgement, to create realistic fake images of real people, or to infer sensitive traits such as health or sexuality. Some laws also prohibit certain practices outright, and those rules differ by country.

Red lines matter because they give staff clarity and protect against gradual drift, where each small step seems reasonable but the end point is not. They also make it easier to say no to a supplier, a client or a senior colleague, because the decision has already been made at the right level.

A red line should be written down in the AI policy, explained with reasons, owned by a senior person and reviewed from time to time, with a clear route to report a suspected breach.

Worked example

A manager at Harbour Foods asks whether an AI tool could read staff's personal social media to predict who might resign.

  1. The team checks the AI policy and finds a red line against monitoring staff's private online activity.
  2. It explains the reason: the intrusion and loss of trust outweigh any planning benefit.
  3. It offers a lawful alternative: anonymous engagement surveys and exit interview themes.
  4. It records the request and the answer in the governance log.

The request is declined quickly and consistently, and the underlying need is met another way.

In practice

Red lines appear in AI policies and supplier contracts, are covered in staff training and are checked during impact assessments. Requests that cross one are declined and logged rather than debated each time.

Watch out

  • A red line is not the same as a high-risk use: high-risk uses may go ahead with safeguards, red lines do not.
  • Red lines set by an organisation can be stricter than the law, which only sets a minimum.

Learn it properly

Official sources

Facts last checked on 3 November 2026.

Education, not advice: the Stockertim Dictionary explains what terms mean. It never recommends buying or selling anything.

English