AI Notice

(Focal Shift™)

Last updated: 3 August 2026

1. Purpose of this notice

Focal Shift uses artificial intelligence systems for part of its services. This notice explains which ones, what they do, what they do not do, and what their limits are.

It is published under the transparency obligations set out in Regulation (EU) 2024/1689 of the European Parliament and of the Council laying down harmonised rules on artificial intelligence (the "AI Act"), and in particular its Article 50.

Following assessment, the systems described below fall under the transparency obligations of that regulation. They do not fall within any of the high-risk categories listed in its Annex III: they are not used for regulated healthcare, for recruitment or employee evaluation, for access to an essential service, credit or insurance, or for educational assessment.

2. The AI systems you may encounter

Every conversation offered on this site is conducted by an artificial intelligence system. None of them is held by a human being in real time.

The AI welcome assistant — available on every page of the site through the chat bubble. It directs visitors to the service or content that fits their situation. It requires neither an account nor identification.
The AI Advisor, in text mode — a paid service available after creating an account. It conducts a structured exchange designed to help clarify a situation.
The AI Advisor, in voice mode — the same service as a spoken conversation. Your voice is automatically transcribed into text, and replies are read back to you by a synthetic voice. That voice is artificial: it does not reproduce the voice of any real person.
The perspective report — a text written automatically from the answers you provide in the guidance questionnaire, then sent to you by email. Its content is AI-generated.
Conversational memory — an optional feature, disabled by default and enabled only by you. When active, an AI system writes a summary of your exchanges so that continuity is preserved from one conversation to the next. That summary, which you can view from your account, is itself AI-generated content.
Automated quality control — exchanges may be subject to an automated evaluation intended to improve the quality of the service. That evaluation concerns how the exchange was conducted, not you as a person, and produces no decision about you.

3. Technical provider

The artificial intelligence models used are provided by OpenAI, acting as a technical processor. The no-retention option is enabled: the data transmitted is neither stored by the provider nor used to train its models.

Details of the data processed, retention periods and legal bases are set out in the privacy policy.

4. What these systems do not do

They make no diagnosis and use no clinical vocabulary.
They issue no prescription and no treatment.
They constitute neither a medical act, nor therapy, nor legal or financial advice.
They make no automated decision producing legal effects concerning you or similarly significantly affecting you.
They do not analyse your emotions by biometric means: no facial recognition and no analysis of your voice or face is carried out. Voice mode is limited to transcribing what you say and reading a reply back to you.
They assign you no score, no rating and no ranking.

5. Limits and possible errors

These systems can produce answers that are inaccurate, incomplete or ill-suited to your situation. This is a known limitation, inherent to the technology, which cannot be entirely eliminated.

An answer may be phrased with confidence and still be wrong. It may also miss an important element of your situation, or offer a reading that does not match you. Treat what you receive as a line of thought, never as established fact — and check for yourself whatever matters to you.

The same applies to voice mode: the transcription of what you say may contain errors, particularly with background noise, rapid speech or a strong accent.

You remain the sole decision-maker for your own choices. Nothing produced by these systems should replace the opinion of a qualified professional when your situation calls for one.

6. Human oversight and reporting

For the sake of transparency: replies are not reviewed by a human being before being sent to you. They are produced and displayed in real time. Individual prior review would be incompatible with how a conversational service works.

Oversight is therefore exercised in other ways:

the operating framework of each system, its safeguards and its prohibitions are defined, tested and maintained by the site publisher;
exchanges are subject to an automated quality evaluation whose results are reviewed so that any drift can be corrected;
every user report is examined individually.

If a reply strikes you as inaccurate, inappropriate or concerning, report it to contact@focalshift.eu. Include the date and the service concerned: your report will be examined and you will receive a reply.

7. Emergency situations

These systems are not designed for emergencies, acute distress or immediate danger, and must not be used in such circumstances.

If you are going through such a situation, reach out to a person without delay:

in France, 3114 — the national suicide prevention line, free, 24/7;
112 — the European emergency number, in case of a life-threatening emergency;
elsewhere, the international directory findahelpline.com.

8. Your rights, contact and updates

You may stop using these services at any time, disable conversational memory from your account, and exercise your rights over your data under the conditions described in the privacy policy. The contractual framework for use is set out in the terms of use.

For any question regarding the use of artificial intelligence on this site: contact@focalshift.eu

This notice is updated as the services and the regulatory framework evolve. The version in force is the one published on the site.