This notice explains when, in using Infinithink, you are interacting with an artificial intelligence system and when content has been generated or altered by AI. It refers to Article 50 of Regulation (EU) 2024/1689 on artificial intelligence (the "AI Act") and to Italian Law No. 132 of 23 September 2025.
1. Purpose of this notice
Infinithink is a platform built on artificial intelligence. We want it to be clear when you are dealing with an AI system and when content has been generated or altered by AI, both for you and for the people who will read, see or hear what you create with the Service.
This notice describes the measures we take and the responsibilities that remain with you. It is in addition to the information shown in the Service interface and supplements the Terms of Service and the Acceptable Use Policy. Capitalised terms such as Service, Input, Output, Customer Content and AI Providers have the meaning given to them in the Terms of Service.
Legal framework at the date of this notice
- the transparency obligations under Article 50 of the AI Act apply from 2 August 2026;
- for AI systems placed on the market before 2 August 2026, the obligation to mark content in a machine-readable format under Article 50(2) applies from 2 December 2026, as provided by Regulation (EU) 2026/1744 (Digital Omnibus on AI), in force since 27 July 2026. The other obligations under Article 50 apply from 2 August 2026 with no transitional period;
- the European Commission has published final guidelines on the transparency obligations under Article 50 and a voluntary code of practice on the transparency of AI-generated content, which the Commission and the European Artificial Intelligence Board have assessed as adequate to cover the obligations under Article 50(2), (4) and (5);
- in Italy, Law No. 132/2025 on artificial intelligence also applies.
2. Who we are and our role
The Service is provided by Lorenzo Locantore, a sole proprietorship under the Italian flat-rate tax regime trading under the Infinithink brand, with business address at Via Marta Tana 19, 46043 Castiglione delle Stiviere (MN), Italy, VAT number IT04570230989, email info@infinithink.it, website infinithink.it, as set out in Section 1 of the Terms of Service. The Service is only available to persons aged 18 or over.
For the purposes of the AI Act:
- we act as the provider of the Infinithink AI system, which we make available under our own brand and which is built on general-purpose AI models developed by third parties such as Anthropic, OpenAI and Google. The AI Providers remain responsible for the obligations that the AI Act places on them as providers of those models;
- business customers, professionals and anyone who uses the Service in the course of an economic or professional activity, including as an individual, act as deployers of the system and have their own obligations, described in Section 6.
If you use the Service solely for personal, non-professional purposes, the deployer obligations under the AI Act do not apply to you. However, if you regularly earn income from what you publish, for example as a creator or influencer, your use falls within an economic activity and you act as a deployer. Even where the law does not require it of you, we nonetheless ask you to indicate that the content is AI-generated or AI-altered when you publish deepfakes or texts on matters of public interest.
3. Which content is generated by AI
All Output produced by the generation features of the Service is generated by AI, in particular:
- texts, articles, emails, social media posts, product descriptions, SEO content and code, including content generated or rewritten with AI in the editor;
- images and, where the feature is included in your plan, videos;
- synthetic voices, voiceovers and other audio content, as well as transcriptions of audio files;
- replies from the chat, the assistants and the chatbots and voice bots built with the Service, including on connected messaging channels;
- content produced by agents and automations, such as blog articles and social media posts, including where it is published automatically on connected websites or accounts.
The AI features use models from AI Providers such as Anthropic (Claude), OpenAI (GPT) and Google (Gemini), which we access either directly or through the OpenRouter routing service. The models available change over time and may differ by feature and by plan. Where a feature lets you choose the model, the model in use is shown in the interface.
4. Interaction with AI systems (Article 50(1))
4.1 The Service is an AI platform, and its chat, assistance and generation features are presented in the interface as AI features.
4.2 Chatbots and voice bots that you build with the Service and connect to websites or messaging channels, such as WhatsApp, Telegram, Instagram or Messenger where available in your plan, display by default, no later than the start of the first interaction, a clear and distinguishable notice informing people that they are interacting with an AI system, in line with applicable accessibility requirements, for example by means of text that screen readers can read or an audible notice. Voice bots announce it at the start of every call. The notice is active on every connected website and channel and cannot be disabled. Where the Service cannot insert it automatically on a channel, we flag this in the interface and you must add it to the welcome message or first reply yourself.
4.3 You may change the wording of the notice, provided it remains clear, easy to understand and accessible, states that an AI system is involved and is given no later than the start of the first interaction. You must not hide or obscure it. In the same message, or in a place that is easy to reach, also provide the link to your privacy notice, which must cover the chatbot (Acceptable Use Policy, Section 9.5). For example, you may use the following wording:
- in English, "You are chatting with an AI assistant. Answers may contain errors."
- in Italian, "Stai parlando con un assistente basato sull'intelligenza artificiale. Le risposte possono contenere errori."
4.4 You must not present a chatbot or voice bot as a person, or configure it so as to lead the people interacting with it to believe that they are talking to a human being.
4.5 The notice is also included in the chatbots and voice bots that we set up on your behalf.
4.6 If you have enabled handover to a human operator, on plans that include it, people can ask to be put in contact with a person.
4.7 If we become aware, including through a report, that a chatbot or voice bot does not inform people that they are interacting with an AI system, we may ask you to correct it and, where necessary, suspend it in accordance with Section 15 of the Terms of Service, with the statement of reasons provided for in Section 15.3 of the Acceptable Use Policy.
5. Marking of AI-generated content (Article 50(2))
5.1 The AI Act requires providers of generative AI systems to ensure that generated content is marked in a machine-readable format and detectable as artificially generated or manipulated, as far as technically feasible and taking into account the specificities of different types of content, the costs of implementation and the generally acknowledged state of the art. As the provider of the Infinithink system, we apply the following measures, as far as technically feasible:
- Images: we apply metadata indicating their artificial origin to image files generated or altered with the Service and, as far as technically feasible, we keep the marks applied by AI Providers even after processing by the Service, such as saving, resizing or format conversion. For example, OpenAI and Google state that they apply content credentials based on the C2PA standard or invisible watermarks such as SynthID to images generated with their models.
- Video and audio: we apply metadata indicating their artificial origin to video and audio files generated or altered with the Service, where the format allows, and, as far as technically feasible, we keep the marks applied by AI Providers.
- Text: in exported files, for example DOCX, PDF and HTML, and in content published through integrations such as WordPress, we insert metadata indicating its artificial origin, as far as technically feasible. These marks are lost if the text is copied as plain text. The visible disclosure required under Section 6 is in addition to the marking and does not replace it.
5.2 We periodically check that the marks remain present in the files produced by the different features of the Service, and we update these measures in line with the state of the art, the European Commission guidelines and the code of practice. At the date of this notice we have not signed the code of practice on the transparency of AI-generated content, and we take its guidance into account for the measures described in this section. We will update this notice when we introduce further measures.
5.3 According to the European Commission guidelines, some Output falls outside the marking obligation, such as source code or short sequences of numbers, symbols or letters. Marking also does not apply to features that only correct spelling and grammar or proofread texts without changing their meaning. Rewriting, rephrasing, translation and image editing with AI, on the other hand, are covered by the marking measures described in this section.
5.4 Machine-readable marks may be altered or lost if content is edited, cropped, converted, compressed or resized with external tools or by the platforms on which you publish it, or if it is copied as plain text or captured in a screenshot. Removing or tampering with marks is prohibited by the Acceptable Use Policy. Marks do not replace the visible disclosures that the deployer must make under Section 6.
5.5 Content credentials can be checked using public verification tools based on the C2PA standard and, where available, the verification tools offered by AI Providers. You can also ask us whether content was generated with Infinithink by writing to info@infinithink.it. We check using the data and tools available to us and reply without disclosing personal data of the user who generated it.
6. Your responsibilities as a deployer
If you use the Service in the course of an economic or professional activity, you are responsible for how you use and publish Output. In particular:
- deepfakes: if you publish or disseminate AI-generated or manipulated images, audio or video that resemble existing or plausibly existing persons, objects, places, entities or events, such as avatars, virtual influencers or synthetic photo models, and that would falsely appear to be authentic or truthful, you must disclose that the content has been artificially generated or manipulated. This includes synthetic voices that imitate real people. You may reproduce the voice or appearance of a real person only with their explicit and verifiable consent, as required by the Acceptable Use Policy (Sections 5, 9.3 and 9.4). Where the content forms part of an evidently artistic, creative, satirical, fictional or analogous work or programme, the disclosure may be made in a way that does not hamper the display or enjoyment of the work (Article 50(4));
- text on matters of public interest: if you publish AI-generated or manipulated text for the purpose of informing the public on matters of public interest, you must disclose that it has been artificially generated or manipulated, unless the text has undergone human review or editorial control and a natural or legal person holds editorial responsibility for its publication (Article 50(4)). Human review means a deliberate examination of the substance of the text by a person with adequate expertise in the subject, including checking facts and sources, and a mere spelling or grammar check is not enough. Matters of public interest include, for example, politics, public health, the environment, consumer safety and economic and financial developments;
- automated publishing: if you set up agents or automations that publish content without your review, for example articles on a WordPress site, social media posts or messages on connected channels, you are responsible towards third parties for that content as if you had published it yourself, unless the publication or its content results from a malfunction of the Service or another breach attributable to us, in which case Section 14 of the Terms of Service applies. Without human review, you cannot rely on the exception for text that has undergone editorial control;
- chatbots and voice bots: you must not hide or alter the AI notice described in Section 4 and, on channels where the Service does not insert it automatically, you must add it yourself on every website and channel where the bot is active. If voice bot calls are recorded or transcribed, you must inform people at the start of the call;
- emotion recognition and biometric categorisation: the Service is not designed or intended for these purposes and must not be used for them;
- professionals: under Article 13 of Italian Law No. 132/2025, members of the intellectual professions may use AI only for instrumental and supporting activities, with their own intellectual work remaining predominant, and must inform their clients about the AI systems used in clear, simple and comprehensive language;
- employers: if you use AI in managing staff, you may have information obligations towards workers under Italian law (Article 11 of Law No. 132/2025 and Article 1-bis of Legislative Decree No. 152 of 26 May 1997). Without our prior written consent, you may not use the Service to select candidates, assess workers or take decisions concerning them, as set out in the Acceptable Use Policy (Section 7).
Disclosures must be clear and distinguishable, conform to the applicable accessibility requirements, be made no later than the first interaction or first exposure to the content, and be perceivable without technical tools, for example by means of a visible label, alternative text or an audible notice. The machine-readable marks described in Section 5 are not sufficient on their own to meet this obligation. For example, you may use the following wording:
- for images, audio and video, "AI-generated or AI-altered content." In Italian, "Contenuto generato o modificato con l'intelligenza artificiale."
- for text, "This text was generated using artificial intelligence." In Italian, "Questo testo è stato generato con l'intelligenza artificiale."
Under Italian law, causing unjust harm to a person by transferring, publishing or otherwise disseminating, without their consent, images, videos or voices that have been falsified or altered using AI systems and are capable of misleading as to their authenticity is a criminal offence (Article 612-quater of the Italian Criminal Code, introduced by Law No. 132/2025).
7. Limitations of AI and human oversight
AI models produce content based on statistical probability, not on verified knowledge. Output may:
- contain factual errors, invented data, quotations or sources, and outdated information;
- reflect biases present in the data on which the models were trained;
- resemble existing content;
- vary in response to the same request.
We recommend that you always review Output before using it, check facts, figures, quotations and sources, and take final decisions yourself, especially in legal, health, financial or reputational matters. The Service is not intended for high-risk uses under the AI Act or for the practices prohibited by Article 5 of the AI Act, as set out in the Acceptable Use Policy.
This information describes the characteristics of AI systems and does not limit the legal guarantee of conformity for digital content and digital services that Consumers enjoy under Articles 135-octies et seq. of the Italian Consumer Code (Section 13 of the Terms of Service), nor Section 14.1 of the Terms of Service.
8. How we select and manage models
- We use models from established AI Providers that publish usage policies and safety measures.
- In addition to the safety filters applied by the AI Providers, we apply usage limits for each plan and may apply further restrictions to prevent uses that breach the law or the Acceptable Use Policy.
- We follow changes announced by the AI Providers, such as the retirement or replacement of models, and update the Service and this notice accordingly.
- We explain how the AI features work and their limitations in this notice and answer questions sent to info@infinithink.it, also to support AI literacy. If you are a business customer, you must take measures to support the development of AI literacy among your staff and other persons who use the Service on your behalf (Article 4 of the AI Act, as amended by Regulation (EU) 2026/1744).
9. Personal data and training
9.1 The processing of personal data when you use the AI features, including the transmission of requests to AI Providers, is described in the Privacy Policy. The Service is hosted on servers located in Germany (Hostinger, Frankfurt). Cloudflare provides the content delivery network and website protection and, where used, file storage. Requests that use an AI model, including text, image, voice and transcription features, are sent to the relevant AI Provider, namely Anthropic, OpenAI or Google, either directly or through OpenRouter, and may be processed outside the European Economic Area subject to the safeguards described in the Privacy Policy. The list of recipients and the safeguards for transfers are set out in the Privacy Policy.
9.2 We do not use Customer Content to train AI models. Content you provide to configure your own tools, such as a chatbot's knowledge base or your brand voice, is used only to operate those tools on your behalf. We use AI Providers through paid services whose terms, at the date of this notice, exclude the use of requests to train models, and through OpenRouter we only make available models from providers that do not use request data for training, as described in the Privacy Policy. If a provider's terms were to change on this point, we would stop using it for the requests concerned. If you use your own API keys for an AI Provider, where your plan allows it, the terms of your contract with that Provider apply.
9.3 Input and Output, including chatbot and voice bot conversations, remain in your account until you delete them or close your account, and are then deleted, including backup copies, within the time limits set out in the Privacy Policy and in Section 16 of the Terms of Service. The same applies to the voice samples and voices you create, where voice features are available in your plan. AI Providers may retain requests for a limited period for security and abuse prevention purposes, under their own terms. For conversations between your chatbots or voice bots and other people, you decide, as data controller, how long to keep them, and you can delete them using the features of the Service, where available, or ask us to do so by writing to info@infinithink.it.
9.4 For the data of people who interact with your chatbots and voice bots, you are the data controller and we act as data processor under the Data Processing Agreement. For the security of the Service and the prevention of abuse, we process some of this data as an independent controller, on the basis of our legitimate interest, as described in the Privacy Policy. Your privacy notice must also cover this processing, for example by referring to our Privacy Policy.
10. Reports and contact
If you find content generated with Infinithink that has been published without the required disclosure, a chatbot or voice bot that does not disclose that it is an AI system, or any other transparency issue, please write to info@infinithink.it, stating where possible where you saw the content. We will assess the report and, where appropriate, take the measures provided for in the Acceptable Use Policy. You can also write to the same address with any question about this notice.
If you believe that your image, your voice or your data have been used with the Service without your consent, write to info@infinithink.it stating the content and where you saw it. We will assess the request and, if it is well founded, take the measures provided for in the Acceptable Use Policy, such as removing the content from the Service or suspending the tool concerned. Reports follow the procedure in Section 17 of the Acceptable Use Policy, and info@infinithink.it is our single point of contact under Regulation (EU) 2022/2065 (Digital Services Act), as stated in Section 18 of the same Policy. You can also lodge a complaint with the Italian National Cybersecurity Agency (Agenzia per la cybersicurezza nazionale), the market surveillance authority for AI in Italy, and with the Italian Data Protection Authority (Garante per la protezione dei dati personali).
11. Updates
The rules on AI transparency are evolving. We will update this notice when there are changes to the law, to the European Commission guidelines and codes of practice, to the available technical solutions or to the Service. The date of the last update is shown at the top of this page. Updates that only concern descriptive information apply from publication. Changes that affect your obligations or rights will be notified to you by email or by a notice in your Account in the manner and within the time limits set out in Section 20 of the Terms of Service, at least 30 days before they take effect, unless the law requires a shorter period. They have no retroactive effect and, if you do not accept them, you may withdraw from the contract without penalty and with a refund of any prepaid amount for the remaining period.
12. Language and governing law
This notice is available in Italian and in English. In the event of any conflict between the two versions, the Italian version prevails, except that for Consumers any doubt of interpretation is resolved in their favour (Article 35 of the Italian Consumer Code).
Matters not covered by this notice are governed by the Terms of Service. This notice is governed by Italian law. If you are a Consumer habitually resident in another country, you keep the protection of the mandatory provisions of the law of that country (Article 6 of Regulation (EC) No 593/2008). Disputes are subject to the rules on jurisdiction in Section 24 of the Terms of Service, including the Consumer's forum (Article 66-bis of the Italian Consumer Code).
Questions about this document? Write to info@infinithink.it.