AI cannot exist without images. Vision requires large quantities of data to learn, test and evaluate models. That's why we build up targeted and varied corpora (BDDs) for our R&D, in full respect of the people concerned and current legislation.
Once created, these databases are annotated and then used for learning and evaluation (performance, errors, bias detection). Our in-house logging and pseudonymization tools enable us to monitor the data lifecycle and protect it from inherent security risks and privacy impacts. We are also working on the augmentation and generation of datasets with synthetic data, free from personal data.
Our dedicated data management team is in constant contact with our partners or customers to support them and raise their awareness of ethics and the RGPD.
As a subcontractor, we provide ongoing support and exchange with the teams concerned for the fulfillment of their legal obligations as data controller relating to the documentation framing the processing of personal data by sending them our processing sheets and carrying out a follow-up analysis.
As for the partners who install our solution, they are made aware of the best practices and challenges of the RGPD and ethics during XXII training and certification courses.
Thanks to our "privacy by design" and "privacy by default" approaches, we can guarantee the best possible protection for individuals. Each feature is designed in collaboration with the product and legal teams. Some of them, with major implications, go through the filter of the ethics committee, which reserves the right to veto development if it deems it necessary. No biometric data analysis, facial recognition, identification or authentication are the cornerstones of our constant commitment to protecting people. This approach is concretely applied to the design for: access rights management, configuration, storage, anonymization of results, API operation and much more.
Our aim is to maintain the optimum balance between the legitimate interests of our customers and the trust of those subject to processing, in our artificial intelligence.
At XXII, we place ethics at the heart of our approach. Transparency, privacy, inclusion, social impact: every decision counts.Find out how we design AI solutions that respect fundamental rights and serve the public interest.
.
The emergence in our societies of technologies such as artificial intelligence demands particular attention, especially from an ethical point of view. Since 2020, an elected ethics committee, made up of members of our teams as well as a panel of external contributors, has been meeting regularly to define and apply the ethical principles that define the development of our product. Ethical issues are fundamental to the design of AI for good. This is why we attach such importance to the ethical use of our CORE, and invite our customers to follow the principles adopted in our charter: beneficence, non-maleficence, autonomy, justice, explicability and transparency. XXII also contributes to the ethics-related work of the ACN (White Paper) and the Hub France IA (Practical Guide).
XXII's product, CORE, is a decision or performance support tool. We use computer vision to complement the human eye, not to replace it. Our software does not lead to any automated procedure following an analysis, it simply facilitates access to and understanding of information already present on video streams, so the use of our product always remains framed by human intervention and respects the values espoused by the CNIL: Necessity lies in the usefulness of the device with regard to the objective pursued, which must be clearly identified. Proportionality resides in the existing presence of video protection or video surveillance. Our product has no impact on the persons concerned in any way other than that provided for by these systems. Data minimization is achieved in particular through the absence of personal data storage by XXII. To date, our software does not store any video streams, and operates in real time. Non-identification: our algorithms consider a person to be an object/silhouette, freeing us from any personal data, so it's the position in space of this silhouette that is analyzed. Finally, the analysis is carried out on a group of people and not arbitrarily. There is therefore no identification or targeted analysis of an individual.
We could not find what you searched for.
Try searching again
Computer vision processes images or video streams containing personal data by incidence. XXII therefore takes care, right from the design stage of its CORE product, to comply with European and French legislation on the processing of personal data, and the XXII teams support their customers in their efforts to protect privacy. New technology and object perception raise new legal, ethical and political issues. As cameras are deployed in both public and private spaces, questions arise about individual freedoms and fundamental rights. Whenever there is a human presence or the presence of an object enabling an indirect association with a person, image capture (photos or videos) and their processing are governed by the RGPD. At XXII, we do not analyze personal data: we distinguish silhouettes that we associate with a category of object (human, dog, car, bicycle...) without ever using biometric data. In addition to object categorization, some of our processes are used for analysis and optimization purposes (assembly line, mobility flow analysis, etc.), and in our algorithmic analyses, a great deal of data, including personal data, can be considered to be "dormant". They are captured by the camera, but do not enter into the artificial intelligence analysis scheme. What's more, our solutions store no personal data and comply with CNIL and RGPD directives and recommendations.
We could not find what you searched for.
Try searching again
AI vision software
Ai vision software