SMEs can use legitimate interest for AI, with a documented test
In brief
On 10 November 2026, the European Data Protection Board adopted final guidelines on legitimate interest under the GDPR. According to the EDPB, this legal basis can support personal data processing for developing and deploying AI models, such as a conversational assistant, fraud detection or cybersecurity, if it is demonstrated. AI is now part of the common doctrine for European supervisory authorities, already accepted in a December 2024 opinion.
What it changes for an SME
For an SME of 10 to 250 people, common AI uses such as a conversational assistant, fraud detection or cybersecurity may rest on legitimate interest rather than consent that is often impractical at scale. You must still demonstrate and document the three conditions of GDPR article 6(1)(f), with a lawful, clear, real and current interest, plus safeguards. A general purpose is not enough: you need the concrete objective of the model. Whether or not the digital omnibus writes this option into the GDPR, supervisory authorities already have a common doctrine for their checks.
What you can do
- State in your records the concrete purpose of each AI model, including the type of fraud or threat you target.
- Document the three-step assessment and the planned safeguards, following the flowchart annexed to the EDPB guidelines.
- Check whether anonymisation or pseudonymisation can reduce the impact on individuals.
Let's talk about the time you could save.
In 20 minutes, we find where AI could save you time or money. Free, no commitment.