In April 2021, the European Commission launched its initial proposal for an Artificial Intelligence Act (AIA). More than three years later, on 1 August 2024, the Act finally entered into force. This chapter revisits the category of ‘prohibited AI practices’ enshrined in Article 5 AIA, which are banned because they are deemed to pose unacceptable risks to the European Union’s values and fundamental rights. After several amendments, the prohibition now covers eight different but interrelated categories of artificial intelligence (AI) practices, namely AI systems that deploy subliminal or purposefully manipulative or deceptive techniques, AI systems that exploit vulnerabilities, AI systems that predict the risk of committing a criminal offence, the creation of facial recognition databases, emotion recognition systems, social scoring systems, biometric categorisation systems and remote real-time biometric identification systems used for the purpose of law enforcement. The chapter examines each of these categories in isolation and critically contrasts the respective changes introduced during the negotiation phase. Last, it undertakes a brief outlook on the interplay between each of these categories and future regulatory challenges in this rapidly evolving field of technology.