At present, there are no comprehensive regulations at the European Union level governing the use of artificial intelligence (AI) in products such as self-driving cars, medical technology, and surveillance systems.
After two years of deliberations, the European Parliament is still debating legislation proposed by the European Commission, which must be adopted by all EU member states before it can enter into force. According to Axel Voss, a German MEP and a key contributor to the EU’s Artificial Intelligence Act, AI has developed significantly during this time, leaving much of the current proposals outdated or obsolete.
It remains uncertain if programs like ChatGPT will be affected by the EU regulations designed to minimize risk associated with AI. The regulation defines these risk levels as ‘unacceptable’, ‘high risk’, ‘limited risk’, and ‘minimal or no risk’ – applications that detect and analyze social behavior for predictive purposes are forbidden. Governments will not be able to employ social scoring, and certain facial recognition technologies have been excluded from the law.
Lawmakers proceed to debate how far AI is permitted to record emotions and simulate them, in addition to assigning categories of risk involved.