Which practice is outright banned under the EU AI Act due to unacceptable risk?

The practice of using AI systems that manipulate human behavior is considered outright banned under the EU AI Act due to its classification as posing an unacceptable risk to fundamental rights and freedoms. This category of AI systems is deemed harmful because they can exploit vulnerabilities in individuals or groups, leading to coercion, manipulation, or social polarization. The EU AI Act specifically targets these types of technologies to protect consumers and society from potential abuse and harmful consequences. By banning these practices, the EU aims to ensure that AI technologies do not undermine individual autonomy or human dignity, reflecting a commitment to ethical standards in AI governance.

In contrast, AI systems used for market analysis, consumer feedback analysis, and assisting in daily administration tasks do not inherently manipulate human behavior to the same degree and are thus considered acceptable under the regulations, provided they comply with other stipulated requirements regarding transparency and accountability.