Artificial intelligence is transforming products, processes and business models. The EU AI Act now establishes binding rules for AI for the first time. Anyone who develops or deploys AI must therefore be able to assess risks, ensure transparency and demonstrate accountability. ePrivacy independently assesses AI use cases against the requirements of the EU AI Act.
The assessment covers everything from risk classification and transparency and information obligations to human oversight and documentation. You receive verifiable proof that your AI application has been designed responsibly and meets the requirements of the EU AI Act.
At the same time, the seal is a visible signal of trust to customers, partners and investors. This helps you navigate the new regulatory framework, strengthen trust in your solutions and turn responsible AI into a competitive advantage.
The AI Act takes a risk-based approach and defines legal as well as technical requirements. These depend on whether an AI system is classified as low-risk, medium-risk, or high-risk. However, some particularly risky applications, referred to as AI systems with systemic risk, are completely prohibited. The evaluation of individuals based on their social behavior (social scoring) and the creation of facial recognition databases through indiscriminate harvesting of facial images from the internet are deemed unacceptable risks and are banned.
High-risk AI systems, such as those used in HR processes or in the medical field, can pose significant dangers to health, safety, or fundamental rights. Therefore, such systems must undergo thorough examination and evaluation before being brought to market. A conformity assessment procedure in accordance with the AI Act is a process through which providers of high-risk AI systems demonstrate that their systems meet the defined requirements before they are marketed or put into operation. Certification can be a potential component of this assessment procedure.
The type of conformity assessment procedure to be conducted depends on the specific category of the high-risk AI system:
Compliance with the AI Act will be mandatory for all organizations that develop, deploy, or use AI systems in the near future. The AI Act is an additional legal framework that exists alongside the GDPR. The GDPR is technology-neutral and applies particularly to AI systems; this will continue to be the case with the AI Act.
