Activities related to the AI Act have reached a crucial moment. After more than several months of discussions, the act has reached the final stage of the process, which is consultation between the European Parliament, the Council and the EU Commission.
The first ever of its kind, the AI Act, was presented by the European Commission on the 21st of April of this year, and on May 11th, the draft negotiating mandate was adopted by the Internal Market Committee and the Civil Liberties Committee, with 84 votes in favor, 7 against, and 12 abstentions. This new draft featured amendments to ensure that AI systems would be overseen by people and are safe, transparent, non-discriminatory, and environmentally friendly. Uniformity is also an important concept in this draft, to ensure the act to be tech-neutral and flexible in applying to the technology of today and of the future.
On 14th of June, the European Parliament met to adopt its negotiating position to initiate the final form of the AI Act draft. An absolute majority of MEPs supported the presented document. The MEPs furthermore expanded the list of AI systems with unacceptable levels of risk to the people’s safety, by banning the use of:
Additionally, the MEPs added AI systems used to influence voters and the outcome of elections to the high-risk list. With all of these regulations, there must be some flexibility left for SMEs to grow their business, and the MEPs added exemptions for research activities and AI components provided under open-source licenses.
The trilogue between the European Parliament, the Council and the EU Commission was held on 18th of July, in which the politicians were to debate controversial sections of the law, mainly concerning innovative measures and so-called regulatory sandboxes.
Regulatory sandboxes are spaces where organizations can experiment with new artificial intelligence tools. This environment is under the control of the proper institution, which benefits by gathering the latest information on particular technological solutions. The main goal is to foster innovation.
Discussions relating to this item were not about its existence itself, but about its mandatory nature. The European Parliament, wanting to make it easier for organizations from smaller EU countries to access the latest techniques, made it mandatory, with the EU Council suggesting that it remain voluntary.
Associated with this law is also the issue of presumption of conformity for authors of AI tools leaving the sandbox. The positions of the Parliament and the Spanish Presidency representing the member states are different. The Parliament’s goal is to provide developers with this compliance, while the Presidency is concerned about the loss of control over the process by regulators, as well as its negative impact on competition.
Generative AI models are also among the key topics discussed in the context of the Artificial Intelligence Act. The configuration of Microsoft’s search tool with the OpenAI model or Google’s much-anticipated Bard has forced appropriate steps to regulate their use to take care of the security of our data.
The new act´s importance should not be overlooked, as it will pave the path for future global legislation on AI. As Brando Benifei, an Italian lawmaker involved in the European Parliament’s act efforts, said,
“Europe has gone ahead and proposed a concrete response to the risks AI is starting to pose. We want AI’s positive potential for creativity and productivity to be harnessed but we will also fight to protect our position and counter dangers to our democracies and freedoms,”
Do you think the act will really set the global benchmark for AI regulation, and do you think this act could be too controlling and limiting?