The voluntary pact on AI signed by Google and OpenAI in the EU

• The European Commission’s proposal seeks to help technology companies adapt to the law on artificial intelligence that recently came into force

The European Commission announced on September 25 that more than 100 technology companies, including Google, Microsoft, OpenAI and Amazon, have signed the voluntary pact that seeks to regulate artificial intelligence (AI) in the countries of the community bloc.

The idea behind this pact is to develop AI that respects the ethical principles and digital laws of the European Union (EU). However, other large companies such as X, Meta or Apple decided not to join the voluntary pact.

In fact, Meta and Apple have halted the introduction of their AI systems into the European market. Both companies have argued that they have doubts about the new laws that the EU has approved to regulate the sector.

“He AI Pact “It is a voluntary instrument, of course, and we invite all companies to participate. The number will increase in the future, but they are private companies and whatever decision they make, it is up to them to do so,” said the spokesman for the Internal Market of the European Executive, Thomas Regnier.

Photo: Pixabay

Brussels promoted the pact not only to promote ethical and responsible development of this technology, but also to help companies adapt to the EU AI lawwhich came into force in August, but whose final deployment will not occur until 2026.

Among the signatories, the European Commission also highlighted companies such as Qualcomm, IBM, the Spanish company Telefónica, Vodafone, Orange, Nokia, Mastercard, Airbus and Booking.

Restrictions on AI in Europe

The European Commission recalled that while promoting this pact, it is also promoting an industrial initiative so that companies can develop their AI systems in the region.

The voluntary agreement will be applied in parallel with the gradual roll-out of the artificial intelligence law, which allows or prohibits the use of technology depending on the risk it poses to people and which will come into force in phases.

Photo: Pixabay

From February 2025, AI systems that categorize people based on biometrics, political, religious, philosophical beliefs, race or sexual orientation will be banned.

Systems that score people based on their behaviour or personal characteristics, or AI capable of manipulating human behaviour, may not be used either.

Also banned will be systems for expanding or creating facial databases collected indiscriminately via the Internet or audiovisual recordings, among others.

In August, transparency criteria that generative AI systems must meet will come into force, which is one of the main points of discussion in the midst of the boom in programs such as ChatGPT.

These models will have to make it clear whether a text, song or photograph has been generated through artificial intelligence and ensure that the data used to train the systems respects copyright.

With information from EFE.

Related news