AI Literacy

AI literacy is the ability to understand how AI systems work, what their limits are, and how to use them knowingly and responsibly. In line with Regulation (EU) 2024/1689 (the AI Act), Qlara is committed to giving every user the information they need to understand and use the platform’s AI features knowledgeably.

The platform embeds artificial intelligence systems in the following areas:

Qlara's AI features are tools that assist with content creation and commercial communication.

Qlara is committed to full transparency about its use of AI:

• AI-generated content is identifiable in the interface
• The user keeps final control: the platform is designed to let you approve content before it is published or sent
• The platform makes it possible to identify chatbot replies as generated by an automated system; it is up to the user to configure and maintain that disclosure (for example in the welcome message or the channel description) so that their own customers are informed

Qlara applies contractual and organisational measures intended to ensure that user data is not used to train third-party AI models. The actual processing arrangements may vary depending on the technology providers used and the specific service configurations. For text generation and chatbot features, Qlara mainly uses open-source artificial intelligence models. For image and video generation, Qlara relies on selected third-party providers, with whom data processing is governed by specific contractual agreements that provide, where applicable, for a ban on using data for model training purposes.

It is important that users are aware of the limits of AI systems:

• AI can generate content that is inaccurate or unsuited to the specific context
• Strategic suggestions are based on general patterns and do not replace professional advice
• Generated images may contain visual artefacts or inaccuracies
• The chatbot may fail to correctly understand complex or ambiguous requests
• AI can reflect biases present in its training data and produce distorted, stereotyped or non-inclusive content: outputs should be checked before they are used or published

Qlara is designed around the "human-in-the-loop" principle: AI assists and proposes, the human keeps control. For content intended for publication or sending (text, images, posts, emails), the platform is designed to let you review and approve the AI output before publication: we encourage you always to check it before making it public. For the chatbot, which operates in real time, the user configures context, instructions and tone in advance and can put escalation paths to a human operator in place. Users can turn AI features off at any time from their account settings.

The AI Act mainly distinguishes two roles. As a rule Qlara acts as the provider and you as the deployer, subject to your own assessment of the specific case. Knowing which role applies to you helps you use the platform compliantly.

What we recommend, in a few points:

• Check before you publish: reread text and images and verify facts, figures, prices and dates
• Be transparent: when an automated assistant replies to your customers, make it clear that it is an automated system
• Do not remove labels: keep the AI-generated content markings where they are present. Removing or altering AI provenance labels, watermarks or metadata may undermine compliance with the transparency obligations under the AI Act
• Protect your contacts' data: process personal data on a valid legal basis and do not enter personal data that is not necessary
• Train the people who work with you: if several people use the platform, make sure they know these basic rules. In particular, we recommend instructing your staff on:
  • what the AI features do and what their limits are;
  • the need always to check generated content before publishing or sending it;
  • the need to be transparent towards their own customers and the prohibition on removing AI-generated content labels;
  • how to handle personal data correctly.
  The materials and guides Qlara makes available can support you in this, but they do not replace the training and organisational measures that remain your responsibility. In any event, meeting the AI literacy obligation and responsibility for the use of the AI features by you and your staff remain yours as the deployer.

We provide practical guides and materials to help you understand and make the most of the AI features of the Qlara platform. To find out more, see the article AI models used by Qlara (which models we use and under which licences) and the guide Using AI knowingly in Qlara in the Help Centre, or contact support.

Can Qlara's AI make up false information?

Yes, it can. Language models can generate content that is plausible but inaccurate (a phenomenon known as "hallucination"). That is why every generated piece of content must be checked by the user before publication. Qlara does not guarantee the factual accuracy of AI-generated text.

Can I use customer conversations or reviews to improve the chatbot?

Any features that learn from conversations, where available, are subject to your informed and active choice: please check and manage the relevant settings in your account. Using content from external sources (for example reviews or comments on third-party platforms) as training material is not recommended; if you choose to upload such material, you remain responsible for the lawfulness of its reuse and for any anonymisation. In any case, avoid entering personal data that is not necessary.

Do I have to tell my customers they are talking to a chatbot?

Yes. The AI Act requires people to be informed when they are interacting with an AI system. Qlara automatically identifies chatbot replies as AI-generated, but we recommend making it clear in the welcome message or in the channel description as well that an automated assistant is active.

Are AI-generated images free of third-party rights?

Not automatically. Qlara does not guarantee that generated images are free of third-party rights; their use is subject to the platform's terms of service and, where applicable, to the terms of the underlying AI providers. AI may also produce output that echoes protected works, styles, trade marks, logos or subjects. Before any use or publication, it is the user's (deployer's) responsibility to verify that the content does not infringe third-party rights (for example copyright, trade marks, image or likeness rights) and, where necessary, to obtain the required permissions. Please also bear in mind that output is not necessarily unique or exclusive: given identical or similar inputs, the system may generate the same or similar results, including for different users. Qlara does not guarantee the exclusivity or originality of generated content.

What happens if the chatbot gives one of my customers a wrong answer?

As the deployer, you are responsible for how the chatbot is used with your customers. We recommend reviewing conversations periodically, configuring the chatbot's context and instructions carefully, and always providing an escalation path to a human operator for more complex requests.

Who is responsible for the content I publish?

Responsibility for content published through mini-sites, landing pages, posts, emails or connected social channels remains yours: you are required to use the platform in compliance with applicable law and to check content before publishing or sending it. Qlara provides the tools, but is not liable for the content you choose to publish.

Which AI models does Qlara use?

For the full list of the models used, their licences and providers, see the dedicated article AI models used by Qlara in the Help Centre.

I want to configure the chatbot or Qlara's other AI features: who can I turn to?

For help configuring AI features (for example setting the chatbot's context, instructions and tone, turning features on or off, or managing generated content), you can consult the dedicated guide Create with Agent Studio in the Help Centre. For more complex configurations, our technical team can give you tailored guidance.