The European Parliament took another step forward in adopting the world's first regulation of artificial intelligence. The move sets off the beginning of a three-way negotiation process between the EU Parliament, the EU Council of Ministers, and the European Commission to finalize the regulation that is likely to take months. The first negotiation meeting will take place as soon as Wednesday evening.
As my colleagues previously discussed, the EU's AI Act takes a risk-based approach to regulations, with an outright ban on applications that are deemed "high risk." The proposed language identifies a number of uses as "high risk," including:
- real-time biometric identification systems, biometric categorization systems that use sensitive data (such as gender, race, ethnicity, religion, and political affiliation);
- predictive policing technology that uses profiling and geolocation; and
- emotion recognition systems in the workplace, schools, and law enforcement.
The law would also ban mass data scraping of facial images from internet or CCTV systems for facial recognition databases.
The next step for the AI Act is the beginning of the negotiations between the lawmakers described above. However, an MEP indicated that they expect these negotiations to be completed by the end of the year. EU member countries will then have some time to set up their regulatory structure and companies will have some time to implement the requirements.