AI NOTICE
Effective date: 2026-03-19 Last updated: 2026-04-19
Preamble: single-operator structure and no Data Protection Officer
Luminibus S.r.l.s. socio unico is a single-person company. The sole director and sole shareholder is the only natural person involved in operating the Service. No Data Protection Officer (DPO / RPD) has been designated, as none of the conditions of Article 37(1) GDPR are triggered at current scale; the internal memorandum recording this determination is available on request at privacy@selv-a.com. The privacy contact point is privacy@selv-a.com; the certified corporate email address (PEC) is luminibus@legalmail.it. Where this notice uses the first person plural (“we”, “us”, “our”), this is linguistic convention for readability and does not imply the existence of additional staff, a dedicated privacy team, a Data Protection Officer, or any third-party personnel beyond the Processors and AI Providers named in the relevant chapter.
1. Purpose
This notice explains how Selv-a uses artificial intelligence, what it can and cannot do, and how your data is involved.
2. What AI does in Selv-a
Selv-a is an AI-based service. Core features rely on AI. AI features include:
- guided reflection flows that adapt to your answers
- generated reflections, reframes, and follow-up questions
- decision-support insights, including comparison of options and devil's-advocate challenges
- personality and psychometric insights, including narrative portraits of how you reason and decide
- mood check-in suggestions and clarity prompts
- comparison insights when you share a reflection with another user (compare and group features)
- automatic titling and short summaries of your flows
- pathway recommendations (Compass) that suggest relevant reflections
- retrieval of your own past content — for example earlier reflections, mood patterns, and similar flows — so that outputs stay consistent with your history
- safety checks that flag concerning patterns and offer support resources.
- Outputs are generated by AI models and are produced without human review, except in the limited cases described in §9.
2-bis. Recommender system disclosure (DSA Article 27)
The Compass feature is a recommender system within the meaning of Regulation (EU) 2022/2065. Its main parameters are:
- Inputs: flows completed in the last 28 days (title, direction tag, timestamp); mood check-ins in the last 7 days (valence, timestamp); user-declared personality type, if any.
- Weights: flow recency is weighted more heavily than flow count; moods older than 7 days are excluded.
- Cadence: Compass is regenerated every Monday at 09:00 UTC and on pull-to-refresh.
- LLM role: a large language model is used to classify the week into directions and to write a short human-readable insight. No decision with legal or similarly significant effects is automated within the meaning of Article 22 GDPR.
- You can turn Compass personalisation off at any time in Settings → Privacy → "Do not personalise Compass". When turned off, Compass returns the same weekly metrics in reverse-chronological order, with no AI classification and no insight text. This is the non-profiling option required by Article 27(3) DSA.
3. AI is not professional advice
AI-generated content in Selv-a is for self-discovery and personal reflection only. It is not medical, psychological, legal, or financial advice. It does not replace a qualified professional. See our Safety & Crisis Notice for emergency guidance.
4. Limitations
AI outputs may be:
- inaccurate, incomplete, or outdated
- inconsistent across sessions
- biased by the data the models were trained on.
- You should not rely on AI outputs for important decisions without independent verification.
5. Which AI providers we use
Selv-a currently routes requests to OpenAI and Anthropic as third-party AI providers. We select providers based on output quality, safety practices, contractual commitments, and data handling posture. If we materially change provider scope or provider terms, we will update this notice and the Privacy Policy before or when the change takes effect.
6. What data is sent to AI providers
To generate outputs, we send the following to the AI provider handling the request, depending on the feature you are using:
- your current prompt and your answer
- selected past-content excerpts retrieved by internal semantic search for continuity — for example earlier reflections, similar flows, and relevant chronicle or outcome entries
- mood summaries, persona summaries, and reasoning summaries generated inside the Service
- language, feature-routing, subscription-tier, and safety metadata needed to handle the request.
- By design, we do not send full account records, payment data, authentication tokens, raw data exports, full audit logs, full support history, or full database dumps to AI providers as part of normal feature requests.
We reduce direct identifiers where feasible, but free-text content may still contain personal or special-category data if you include it.
7. Third-party model training and retention
Under Selv-a's current commercial setup, inputs and outputs sent to OpenAI and Anthropic are not used to improve provider models unless Selv-a explicitly enables a separate feedback or sharing setting.
Data sent to AI providers is still processed to generate outputs and may be retained under the provider's platform settings, endpoint behavior, abuse-monitoring rules, and legal obligations. Selv-a does not currently claim OpenAI Zero Data Retention or Modified Abuse Monitoring.
8. How Selv-a uses your data internally
Selv-a does not currently use your entries or interactions for non-essential internal AI or service-improvement activities such as prompt tuning, model evaluation, or provider benchmarking.
We process content where necessary to provide the requested feature, maintain safety and security, investigate abuse or incidents, troubleshoot technical problems, or comply with law.
9. Human review
Authorised personnel may review content only where necessary for safety, abuse, security, technical troubleshooting, legal compliance, or to establish, exercise, or defend legal claims, under confidentiality and access controls.
10. Your choices
You can:
- avoid including sensitive personal information in your inputs
- grant or withdraw the optional AI-related consents (personality profiling, mood tracking, AI insights) at any time in Settings, without losing access to your account or the rest of the Service
- report an AI output that feels unsafe, inaccurate, or inappropriate — reports are reviewed under our AI Safety Incident Workflow
- export or delete your data
- exercise your rights under GDPR, including access, rectification, restriction, objection, and data portability, as set out in the Privacy Policy
- manage marketing preferences
- submit privacy requests, including restriction requests and objections, to privacy@selv-a.com
- delete your account at any time
- lodge a complaint with a supervisory authority — for users in Italy, the Garante per la Protezione dei Dati Personali.
11. Sensitive reflections
Because Selv-a is a self-discovery service, your reflections may include personal topics (for example, mood, stress, or life concerns). Such information may qualify as special category data under GDPR Article 9 and receives additional protections. See our Privacy Policy Chapter 4.5 for how we handle it.
12. Contact
Questions about AI in Selv-a: legal@selv-a.com Privacy requests: privacy@selv-a.com
13. Related policies
Privacy Policy: https://selv-a.com/privacy Terms and Conditions: https://selv-a.com/terms Safety & Crisis Notice: https://selv-a.com/safety-notice