Members of the European Parliament debated key provisions of the upcoming artificial intelligence regulatory bill for 15 hours yesterday. In the end, officials managed to agree on a number of key issues, paving the way for further progress on the bill.According to Reuters, the list of approved rules is quite extensive, but several main points can be highlighted. Therefore, the upcoming law will impose an obligation on companies that provide the underlying models to ensure the transparency of the technology and the disclosure of the data on which the neural network is trained. It is stated that this is necessary to comply with EU copyright law.
Some regulations received criticism
The law also stipulates that the government can use real-time biometric surveillance for citizens in public places. However, it is claimed that this surveillance will only be used to record crimes and search for actual or foreseeable threats to public safety and suspected criminals. At the same time, such social surveillance systems were prohibited from performing biometric categorization based on political, religious and philosophical views, gender, race and other considerations.
European human rights activists and IT companies have already criticized the EU’s first rules on artificial intelligence. DigitalEurope, for example, called the attempt to regulate AI “another burden for companies.” European Digital Rights is concerned about the de facto legalization of real-time facial recognition.The first implementation of EU Artificial Intelligence regulatory law could come into force as early as next year.
36225
