Content moderation policy

Article 1 — Purpose and scope

1.1. This content moderation and acceptable use policy (the "Policy") defines the permitted and prohibited uses of the artificial intelligence services provided by LTC Group SRL ("LTC"), as well as the monitoring, investigation and enforcement practices implemented by LTC to prevent and handle abuse.

1.2. It applies to all AI Services, including in particular: (i) AI Solutions designed, configured or operated by LTC for its clients (business AI agents, assistants, chatbots, voice agents, automations, document generation, analysis and classification); (ii) software products published by LTC incorporating AI (Nota / Sérénotaire, Dedal IA, Léa, LTC Demo Studio…); (iii) free AI tools on the Website, including the AI solutions generator; (iv) demonstration, trial and training environments.

1.3. The Policy forms an integral part of the terms of use and the general terms and conditions of sale (in particular Article 13 "Artificial intelligence: specific rules"). It is binding on the Client, which shall ensure compliance by its Users, as well as on any person interacting with an AI Service. In the event of a conflict, the provision offering the greatest protection for third parties and legal compliance shall prevail. The terms specific to an LTC Product may supplement it.

Article 2 — Definitions

AI Services

Any feature provided, developed, hosted or operated by LTC that relies, in whole or in part, on an artificial intelligence model, as described in Article 1.2.

Client

The natural or legal person ordering or subscribing to an AI Service from LTC, in its capacity as deployer within the meaning of Regulation (EU) 2024/1689 ("AI Act").

User

Any person using an AI Service: the Client's staff, visitors to the Website, as well as the Client's customers, patients, prospects or contacts who interact with a chatbot or voice agent.

Content

Any data submitted to an AI Service (instructions, questions, files, recordings, "prompts") or produced by it (texts, replies, summaries, voice, images, automated actions).

Breach

Any use or Content contrary to this Policy, the Terms of Use, the GTC, a model provider's terms or the law.

Article 3 — Permitted uses

3.1. The AI Services are intended for lawful, fair and transparent professional uses within the scope of the Client's business activities. The following are permitted in particular:

• automation of administrative and commercial tasks (quotes, invoices, follow-ups, emails, data entry, synchronisation between software tools, appointment booking);
• assistance with drafting, translation, summarisation, classification and information retrieval in the Client's documents;
• customer support and reception via chatbot or voice agent, provided that individuals are informed that they are interacting with an AI and can reach a human;
• internal data analysis and reporting, preparation of decisions subject to human validation;
• creation of honest marketing content, reviewed by the Client before publication;
• training, demonstration and experimentation in a controlled environment.

3.2. Any permitted use remains subject to applicable law (in particular GDPR, AI Act, Code of Economic Law, copyright, labour law), to safeguards agreed with LTC and to human oversight proportionate to the stakes.

Article 4 — Prohibited uses

4.1. It is strictly prohibited to use an AI Service, attempt to use it or allow a third party to use it for:

1. Unlawful content: producing, storing or distributing any content contrary to the law, and in particular any child sexual abuse material (CSAM) or content sexualising minors, without any exception;
2. Violence and extremism: inciting, promoting or glorifying violence, terrorism or self-harm, or providing assistance in planning violent acts;
3. Hate and harassment: producing hateful, discriminatory, threatening, abusive or harassing content targeting an individual or group, particularly on the basis of origin, religion, sex, sexual orientation, disability or age;
4. Weapons and dangerous materials: obtaining assistance in the design, manufacture or acquisition of weapons, explosives or chemical, biological, radiological or nuclear agents;
5. Fraud and deception: scams, phishing, fake reviews, fake invoices, fake documents, identity theft, market manipulation, as well as creating deepfakes of text, voice or images of a real person without their express consent;
6. Disinformation: generating or distributing false or misleading information on a large scale, including in electoral, health or financial matters, or passing off AI-generated content as human content where the law requires disclosure;
7. Spam and unlawful direct marketing: mass unsolicited communications, automated calls to individuals registered on the "Do Not Call Me" list or without a legal basis, misleading robocalls;
8. Invasions of privacy: collecting, profiling or processing personal data without a legal basis, compiling databases of individuals without their knowledge, identifying an individual from their voice or face, or disproportionately monitoring workers, clients or acquaintances;
9. Prohibited practices under the AI Act (Article 5): subliminal or manipulative techniques, exploitation of vulnerabilities linked to age, disability or social situation, social scoring, individual risk assessments of offending based solely on profiling, emotion recognition in the workplace or education, sensitive biometric categorisation, untargeted scraping for facial recognition databases;
10. Security breaches: creating malware, discovering or exploiting vulnerabilities without authorisation, circumventing security measures, conducting cyberattacks;
11. Circumvention of safeguards: attempting to disable, evade or circumvent filters, system instructions, rate limits or moderation mechanisms of an AI Service (prompt injection, "jailbreaking"), or extracting instructions, other clients' data or confidential secrets;
12. Infringement of third-party rights: infringement of copyright, trademarks or trade secrets, imitation of the voice or style of an identifiable person without authorisation;
13. Sexually explicit content or pornographic material;
14. Unsupervised automated decisions producing legal or similarly significant effects on an individual (refusal of credit, employment, healthcare, housing, contract) without effective human intervention;
15. Unfair competition and service abuse: reselling access to an AI Service without agreement, using it to train a competing model, deliberately overloading it or making manifestly disproportionate use of it.

4.2. These prohibitions apply to submitted Content as well as produced Content, regardless of the channel (web interface, API, telephone, messaging, automation) and even when the request is presented as fictional, humorous, testing or research-oriented.

Article 5 — Restricted, sensitive or high-risk uses

5.1. The following uses are not prohibited in principle, but are only permitted with LTC's prior written consent, a documented risk assessment and enhanced safeguards agreed in the Quote:

• health and well-being: any use related to patients, symptoms, treatments or health data;
• regulated professions: legal, notarial, tax, accounting, financial, insurance or investment advisory uses;
• employment and human resources: CV screening, employee assessment, management or monitoring;
• education: assessment, guidance or admission of pupils or students;
• minors: any AI Service intended for persons under 18 years of age or likely to be predominantly used by them;
• sensitive data within the meaning of Article 9 GDPR (health, beliefs, religion, sexual orientation, biometric or genetic data) and judicial data;
• voice agents making outbound calls, recording conversations or using cloned voices;
• autonomous actions: agents authorised to send, pay, order, delete or alter data with third parties without human validation;
• political, electoral or public interest content intended for public distribution;
• public services, critical infrastructure, access to essential services and, more generally, any use falling under high-risk AI systems in Annex III of the AI Act.

5.2. For these uses, LTC may require in particular: systematic human validation before any external effect, enhanced disclaimers, exclusion of specific topics, data pseudonymisation, European or local hosting, extensive logging, a data protection impact assessment (DPIA) conducted by the Client, prior testing and periodic review. LTC may refuse a sensitive or high-risk use without having to provide justification.

5.3. AI Services never replace a healthcare professional, lawyer, notary, certified accountant or financial advisor. They must never be used in life-threatening emergency situations: in case of emergency, call 112.

Article 6 — Obligations of the Client and Users

6.1. The Client shall: (i) use the AI Services in accordance with their documentation, this Policy and the law; (ii) inform its Users and data subjects that they are interacting with an AI and that generated content may contain errors; (iii) organise human review proportionate to the stakes before any distribution or decision; (iv) hold a legal basis for the data it submits and only provide necessary data; (v) ensure AI literacy of its Users (Article 4 of the AI Act); (vi) promptly report to LTC any incident, anomalous behaviour, inappropriate content or abusive use of which it becomes aware; (vii) cooperate in good faith with any investigation conducted under Article 9.

6.2. The Client is responsible for any use made of the AI Services from its accounts, access credentials and channels (website, telephone number, messaging), including by its Users and its own customers. The Client indemnifies LTC against any claim resulting from a Breach attributable to it.

Article 7 — Built-in safeguards in our services

7.1. LTC designs its AI Services to reduce the risk of abuse by design. Depending on the nature of the service, these safeguards include in particular: system instructions limiting the agent to its business scope and prohibiting out-of-scope topics; use of model providers' safety and moderation filters; additional input and output filters (keywords, risk categories, detection of circumvention attempts); explicit disclosure of the artificial nature of the interlocutor; escalation to a human; mandatory human validation before sensitive actions; rate, volume and consumption limits; data minimisation or pseudonymisation.

7.2. As no safeguard is infallible, these measures do not relieve either the Client or Users of their own obligations.

Article 8 — Monitoring and detection

8.1. LTC reserves the right to monitor the operation and use of the AI Services in order to detect, prevent and address Breaches, security incidents and malfunctions. This monitoring relies in particular on: technical logs (timestamps, volumes, errors, triggered filters); automated alerts in case of abnormal volumes, repeated circumvention attempts or content flagged by filters; sample audits as part of agreed maintenance; reports received under Article 11; notices from model providers.

8.2. Human access to conversation content is strictly limited to what is necessary for maintenance, security, handling a report or reasonable suspicion of a Breach, or complying with a legal obligation. It is restricted to authorised LTC personnel bound by confidentiality, and documented. LTC does not perform systematic reading of conversations and does not use Client data to train models (Article 13.5 of the GTC).

8.3. These processing operations are based on the legitimate interests of LTC and the Client in ensuring the security and lawfulness of the services, and, where applicable, on compliance with legal obligations. Their retention period and the rights of data subjects are described in the privacy policy and, for AI Solutions operated on behalf of a Client, in the data processing agreement (DPA).

Article 9 — Investigation and handling of abuse

9.1. When a Breach is detected, reported or reasonably suspected, LTC may investigate. To this end, LTC may in particular: examine logs and relevant Content within the limits of Article 8.2; request explanations, information or documents from the Client, to be provided within the specified deadline; temporarily suspend the relevant feature as a precautionary measure during the investigation; preserve evidence for the necessary duration.

9.2. LTC handles reports diligently, objectively and in a non-arbitrary manner. Urgent reports (threat to life or personal safety, child sexual abuse material, active fraud, security breach) are handled with top priority.

9.3. LTC may inform the competent authorities (police, public prosecutor's office, Data Protection Authority, Centre for Cybersecurity Belgium, Child Focus) of any facts likely to constitute an offence, cooperates with them and provides them with information required by law. LTC may also inform the relevant model provider where required by its terms.

Article 10 — Measures in case of breach

10.1. In case of a Breach, LTC may take one or more of the following measures, individually or cumulatively:

1. a warning addressed to the Client, requesting compliance within a specified timeframe;
2. the refusal, blocking or removal of Content, a request or a response;
3. the strengthening of safeguards or reconfiguration of the AI Service (excluded topics, mandatory human validation, limits);
4. the restriction of access: rate or volume limitation, deactivation of a feature, channel (number, widget, integration) or User;
5. the temporary suspension of the relevant account, access or AI Service;
6. the termination of the service or contract, under the conditions of Article 21 of the GTC;
7. reporting to competent authorities and preservation of evidence;
8. any appropriate legal action, in particular seeking compensation for damages incurred.

10.2. Measures are proportionate to the severity, repetition and intentionality of the Breach as well as the risk to third parties. In principle, LTC issues a warning first. However, in case of a serious Breach (in particular Article 4.1, points 1, 2, 4, 5, 9 and 10), emergency, risk to personal safety or system security, requirement by an authority or model provider, or repeat violation, LTC may suspend or terminate immediately without prior notice.

10.3. A suspension or termination motivated by a Breach attributable to the Client entitles the Client to no indemnity or refund of sums paid or due, without prejudice to mandatory consumer rights. Reasonable expenses incurred by LTC to investigate and remedy the Breach (investigation, remediation, excessive model consumption) may be reinvoiced. LTC may lift the measure once the Breach has ceased and adequate guarantees have been provided.

Article 11 — Reporting abusive content or use

11.1. Any person, client or not, may report Content or use of an AI Service that they deem unlawful or contrary to this Policy, by writing to contact@ltcai.be with the subject line "Report – AI moderation".

11.2. To enable effective processing, reports should preferably include: a description of the Content or use concerned and the reasons why it is considered unlawful or abusive; its precise location (website, telephone number, date and time, screenshot or extract); contact details of the reporting person (except for child sexual abuse material, which may be reported anonymously, including to Child Focus via stopchildporno.be); a statement confirming that the information is accurate and submitted in good faith.

11.3. LTC acknowledges receipt of the report, processes it within a reasonable timeframe and informs the reporting person of the action taken where contact details were provided. Manifestly unfounded or abusive reports may be disregarded. Security vulnerabilities are reported in accordance with Article 8 of the Terms of Use.

Article 12 — Reasoning, contestation and appeals

12.1. When LTC restricts, suspends or terminates an AI Service or removes Content, it informs the affected Client no later than when the measure takes effect, unless prohibited by law, an authority or a serious risk. The notice specifies the measure taken, its duration, the factual circumstances and grounds on which it is based, and available remedies.

12.2. The Client or reporting person may contest a decision within six months by writing to contact@ltcai.be with any relevant information. The contestation is reviewed by an LTC representative who was not involved in the initial decision, and a reasoned response is provided within a reasonable timeframe. The measure may be lifted or adjusted if found to be unjustified.

12.3. This internal appeal procedure does not deprive the person concerned of the right to initiate proceedings before the competent courts, the Data Protection Authority or, for consumers, the Consumer Mediation Service, in accordance with Article 25 of the GTC.

Article 13 — Model providers' policies

13.1. The AI Services rely on models provided by third parties (in particular Anthropic, OpenAI, Google, Mistral, ElevenLabs) or on open models. Their acceptable use policies apply in addition to this Policy; the stricter rule shall prevail.

13.2. A provider may, on its own initiative, filter Content, refuse a request or suspend access in the event of a breach of its rules. LTC is not liable for such decisions, but will inform the Client and assist in restoring service where possible.

Article 14 — Amendments and contact

14.1. LTC may amend this Policy to reflect changes in its services, observed risks, requirements of model providers, payment partners or regulations. The applicable version is that published on the Website; its date and version number appear at the top of the page. Clients under contract are notified of substantial changes.

14.2. This Policy is governed by Belgian law. It is drafted in French; its translations are provided for convenience and the French version prevails.

14.3. Contact: LTC Group SRL, Rue du Noir Jambon 13, 7830 Thoricourt (Silly), Belgium — contact@ltcai.be — +32 466 43 16 55.

---

See also: terms of use, terms and conditions of sale, privacy policy, legal notice, GDPR & AI Act compliance.