L'AI Act, the European regulation that will regulate artificial intelligence in the Union, will come into force shortly but will be fully applicable in just over two years. In the meantime, the Commission has launched an AI Pact, i.e. a pact that encourages companies to prepare in advance for the new rules. How should they do it? “Without anxiety, with great competence and awareness” he tells Adnkronos the lawyer Licia Garotti, partner of the PedersoliGattai law firm and expert in Technology and Industrial Property law. “Every reality should ask itself two questions: am I already using AI technologies? Often they are technologies already implemented in business processes, but unconsciously. Are we ready to make the best of it? Without stopping at generative AI alone, the different forms of artificial intelligence bring very important advantages for productivity, innovation and, therefore, competitiveness on the markets. AI technologies must also be addressed in a context that is not only regulatory (necessary, but not sufficient). It is made up of different threads: technology, industry and governance, ethics (essential), regulation, economic-social response, impacts on sustainability. Without the right threads, there is a risk of holes and voids. And this is precisely the right time for a reasoned organization (and protection) of data and processes”, explains Garotti.
A question that arises, especially in the entrepreneurial field, concerns the braking effect that the regulation could have on innovation. “The most serious mistake for European companies and, first and foremost, Italian ones, is to experience the adaptation to the AI Act as a “plaster”, a block to development and operations”, explains the lawyer. “The Regulation can actually represent an effective guide for developers and users. This is done by strategically exploiting, for example, the AI Pact. It is a system of anticipated and voluntary compliance (Anticipated Voluntary Compliance) which starts from the risk impact assessment. With an early adaptation to the rules already well outlined in the version approved by Coreper on 2 February, European companies have the opportunity to embrace the opportunities of continuously evolving technology”.
The regulatory framework does not end with the AI Act: it is also necessary to consider the directive on liability from AI systems (AI Liability Directive), the digital service package (Digital Markets Act and Digital Services Act), national laws and, depending of the sector, the different reference regulations. Among these, lawyer Garotti cites “the Dora Regulation aimed at financial and insurance services for the security and resilience of digital infrastructures in Europe; or the regulation of artificial intelligence in European legislation on medical devices”. Among the positive elements of having a law that applies to the entire Union is standardization: “Various organizations operate in this field that try to standardize processes for all Europeans, reducing costs and bureaucratic difficulties. Among these is CEN-CENELEC which, through the CEN-CENELEC Joint Technical Committee 21 dedicated to AI, is responsible for the development and adoption of standards for AI and related data, as well as providing guidance to other committees technicians involved. At the same time, European and Italian legislation plays a decisive role on the projects, initiatives and incentives adopted to encourage investments in the production of semiconductors and chips. In this context, Italy is playing a primary role in preparing suitable measures to attract investments – including from foreign entities – in the semiconductor and microelectronics sector”, explains lawyer Licia Garotti to Adnkronos.
For some, more space has been given to competitiveness – so as not to fall too far behind China and the United States which currently do not have legislation on the matter – than to fundamental rights. For Garonna we need to look not to Brussels but to Strasbourg, seat of the Council of Europe, and New York, in the United Nations building: “The draft framework convention on artificial intelligence, human rights will also try to give an answer on rights , democracy and the rule of law, published by the Committee on Artificial Intelligence (CAI) of the Council of Europe in Strasbourg on 14 March 2024 (still in confidential form, although not officially available). Parties shall take appropriate legislative, administrative or other measures to implement the provisions contained in the Convention, which may include specific or horizontal measures regardless of the type of technology used. The latest resolution of the United Nations General Assembly on the promotion of “safe, secure and trustworthy” artificial intelligence (AI) systems, for the benefit of sustainable development for all, also fits into this framework. However, the EU is still behind on investments (5-10 times lower than the United States, often without coordination between the countries of the Union)”, concludes Garotti.
#Italian #companies #prepare #legislation