Which area does not fall under high-risk AI systems according to the EU AI Act?

The category of AI systems that does not fall under the high-risk designation according to the EU AI Act is related to spam filtering. The EU AI Act classifies AI applications into different risk tiers, with high-risk applications subject to stringent regulations and oversight.

High-risk AI systems are defined as those that pose significant risks to the rights and safety of individuals, and they include areas such as biometric identification, which can impact privacy and security, and the management of critical infrastructure, where failures could have dire consequences for public safety. Similarly, education and vocational training can also fall under high-risk categories as these systems can impact individuals' future opportunities and personal development.

In contrast, spam filtering is generally considered a low-risk application. While it plays an important role in managing unsolicited communications, it does not involve critical safety considerations or significant impacts on individual rights in the same way that high-risk categories do. The regulatory framework prioritizes oversight based on the potential implications of how these technologies are implemented and their potential impact on society. Therefore, spam filtering is classified outside of the high-risk category according to the EU AI Act.