Which of the following is a high-risk AI system that requires strict regulatory compliance?

Facial recognition in public spaces is considered a high-risk AI system primarily because of the significant implications it has on individual privacy and civil liberties, as well as its potential for misuse or bias. Such systems can track individuals without their consent, leading to concerns surrounding surveillance and the erosion of anonymity in public life.

Regulatory compliance is essential for these systems due to the ethical considerations and potential impact on society. Laws and regulations often require high levels of accountability and transparency for technologies that can affect fundamental rights. This is especially crucial given the tendency for facial recognition algorithms to exhibit biases based on race, gender, or age, which raises important questions about fairness and discrimination.

In contrast, while virtual assistants, spam filters, and social media recommendation systems may have their own regulatory concerns, they typically do not pose the same level of societal risk and controversy associated with facial recognition technology. Therefore, the strict regulatory compliance is most relevant to the deployment and use of facial recognition in public spaces, distinguishing it as a high-risk AI system.