Scriptor Artis is committed to protecting your data. From encryption to processing transparency, we're committed to giving you the best possible insight into how Devana works and how data is processed by its model providers. All our tools are designed for you, and you alone.
To guarantee data confidentiality, Scriptor Artis applies rigorous encryption to all data and communications. This means that all information sent and received is encrypted so that it can only be read by authorized parties. For further information, please consultour privacy policy.
All conversations are end-to-end encrypted.
Your knowledge base data is also stored with our hosting partnerScalewayequipped with reinforced security systems. For added security, we replicate your backups on 3 remote servers and perform cold storage once a day. Our partner has also been awarded the SecNumCloud qualification visa for 2019. This makes him a trusted player in the French cloud. We also offer the following servicesOn-Premise to host Devana in your home. Please do not hesitate to contact the Scriptor Artis team.
At Scriptor Artis, we never use your data to train or process models. We've put in place advanced systems that allow you to direct your searches. From personalizing your agent to giving it access to selected knowledge bases, you're in control of the information you pass on to your agent. To find out more about your agent's behavior, read our explanations of embedding and fine-tuning. All our systems can be parameterized by you.
In the event of a web query, your sensitive data is neutralized.
When you use Devana to search the web, your data is systematically cleaned so that you can formulate neutral search engine queries.
We monitor the data processing conditions of our model suppliers on a daily basis. You can consult our reports onnext link.In the event of an update that does not comply with regulations or that could compromise the security of your data, we reserve the right to suspend access to a model. You will be offered a replacement model.
We keep you informed of any updates from our suppliers.
All updates and data processing policies of our model providers are reported in the information points of the Devana application. They remain accessible via this link.
Scriptor Artis has implemented moderation systems to reduce dangerous or sensitive content. However, it is advisable to test your agents before sharing them. In fact, your knowledge bases may transmit inappropriate or sensitive information. Furthermore, the identity you give your agent may be inappropriate. You remain the master of your agents.
In the event of an inappropriate response, you have access to the reporting tool at any time.
Scriptor Artis also warns its users about answers provided by Devana. We provide our users with tools for verifying sources. These tools are designed to help you develop a critical mind. It remains imperative that content be checked by the user if he or she wishes to ensure the conformity of a piece of information. If you wish to strictly control your agents, we recommend choosing the limited AI mode. This mode requires your agent to rely solely on the knowledge bases provided.
Scriptor Artis provides you with storage space to build up your knowledge bases. From these bases, you can build intelligent agents for yourself or for your customers and collaborators. We would like to warn agent designers about making available copyrighted sources.
While Scriptor Artis provides tools for citing sources and verifying sources, it is the responsibility of designers and users to ensure that they have the right to use sources and to cite their use. Our responsibility lies in providing ethical tools, and the user's in using the sources. Every designer must check the legislation in force concerning copyright before building up knowledge bases, particularly those shared in public.
Scriptor Artis applies current regulations in the provision of its services. In the interests of transparency with all our users, we regularly share news about AI and its regulation. Find our posts on this subject on ourblogblog. However, we invite our readers and users to keep their own watch on the subject, by consulting the CNIL website, for example, or the European Commission's website dedicated to the AI Act (Artificial Intelligence Act), the draft of which has yet to be voted on definitively.
We are committed to respecting the integrity of our users.
While the AI Act has not yet been passed, we have made strong commitments to ensure that AI remains an ally and does not become a threat. We assure our users that our tools have no vocation to be authoritative over them, to manipulate them psychologically or to establish a social score. We remind you of the responsibility of our intelligent agent designers. Every user has the right to report any inappropriate behavior on the part of an agent at any time.
Forging our critical thinking
Generating progress
