Europe has taken a crucial step to regulate AI. This Wednesday, The European Parliament approved by a large majority its final position on the Artificial Intelligence Law. Once this has been achieved, the next step will be negotiations with the European Commission and the member countries of the European Union to give final form to the legislation.
The vote in the Eurocámara culminated with 499 votes in favour, 28 against and 93 abstentions. In any case, the final position on the Artificial Intelligence Law does not mean that the regulation is ready to be adopted. The discussions with the so-called trialogues will be crucial to obtain a final version of the regulatory text, which is expected to come into effect no later than 2026.
As reported luca bertuzziof euractiv, no last-minute changes were made to the text adopted at committee level. This implies, finally, that the amendment proposed by the European People’s Party, which sought to allow the use of facial recognition technologies and other biometric systems by law enforcement in real time, has been rejected.
It is important to clarify, however, that the implementation of Remote Biometric Identification has not been completely excluded from the Artificial Intelligence Law. MEPs have approved its use in post-serious crime investigationswith a court order involved.
What the Artificial Intelligence Law says about ChatGPT and other generative AI
The position of the European Parliament on the Artificial Intelligence Law that ratifies the implementation of a risk categorization system. This means that specific regulation will not apply to the technologies that power generative AI platforms, but rather to how they are used.
This has already generated controversy with figures like Sam Altman. It is that the legislation that the European Parliament supports, places ChatGPT and similar tools in the “high risk” category. That is to say, in the second highest and only behind the artificial intelligence systems that will be prohibited as unacceptable.
In fact, the CEO of OpenAI himself stated a short time ago that he did not object to European efforts to regulate AI. But that he was against the way the Artificial Intelligence Law was written. The truth is that, under the text that today has the permission of the European Parliament, technology providers must meet a series of strict requirements.
The authorities intend that the language and foundational models be designed so that cannot create illegal content, and to identify what copyrighted materials are used to train them. Let’s not forget that, at the time, it was Altman who threatened to leave Europe if ChatGPT and other products from his startup could not comply with the new Artificial Intelligence Law. “If we can deliver, we will, and if we can’t, we’ll go out of business. We’ll try. But there are technical limits to what’s possible,” he said during a recent UK appearance.
The European Parliament has today taken a crucial step to negotiate the world’s first Artificial Intelligence Law. However, there is still a long way to go and the discussions with the EU member countries and the European Commission promise to be extensive. Negotiations are expected to intensify once Spain assumes the presidency of the Council of the European Union on July 1.