European Commissioner wants AI code of conduct for tech companies within a few weeks

European Commissioner Margrethe Vestager wants the US and EU to encourage major tech companies to draw up a voluntary code of conduct for AI within a few weeks. This should serve as a temporary measure while work is being done on a binding Artificial Intelligence Act.

Vestager says during a press conference in Sweden that such a code of conduct could make people “more comfortable” with the fact that generative AI exists and is developing rapidly. reports Reuters. She wants the EU and US to take a joint lead and ensure that agreements are made with tech companies on such a code. There should be a draft version of such a code of conduct within a few weeks; the final version could then be ready ‘very, very quickly’. “You need that kind of speed if you want society to trust current developments.”

Although work is already underway on European AI legislation, according to the European Commissioner it will not come into force in two and a half to three years at the earliest. “That is clearly too late. We must act now.” She does not say what must be agreed in that code of conduct, but it must contain agreements ‘on details, not just general statements’. Google and European Commissioner Thierry Breton previously indicated that they were working on an AI pact.

The European Artificial Intelligence Act was tabled in 2021 and was approved by the telecom ministers of the EU member states in 2022. If it comes into effect, it could be the first comprehensive AI law. This law must make it clear when an AI system is considered safe and establish a number of criteria for this. This should allow AIs to be classified into three levels: unacceptable risk, high risk and limited risk. The first AI level has been banned in the EU after the introduction of the law; systems that meet the second level are only permitted under strict conditions.