Many companies are holding their breath ahead of Parliament’s vote on the AI Act. And maybe many of us should do the same?
Much is at stake for how we will be able to take advantage of AI in Europe in the future. From a company perspective, it is about competitiveness, market shares and job opportunities. The AI proposal has been under discussion for months in Parliament and initially many relevant and important improvements were made to the Commission’s proposal of April 2021. In recent weeks, however, the changes have become so extensive that they are difficult to interpret and foresee the consequences thereof.
Some companies insist that the administrative burdens for using so-called high-risk AI and generative AI result in a delay of up to two years before a new product or service can be put on the market. It does not work in practice. Then Europe will have lost rapid innovation and the use of beneficial new technologies. It would affect the entire data economy.
In addition to the AI Act, the Data Act is also negotiated. The final version of these two regulations will affect business models and business opportunities in a very extensive way and as it stands now, parliamentarians cannot possibly have had time to complete that analysis. More time and (AI) knowledge is needed for the regulations to end up right and benefit both business and society at large.
It would be worth slowing down the process and rethinking. It will be difficult to include all the different definitions and techniques, especially since research is advancing so rapidly. New things have come out before the former have even been legislated for.
In the best of worlds, as the proposal stands now, we should skip the different definitions and focus more on uses, such as Annex III, Article 4a (general principles for trustworthy AI), Article 5, Article 6 and Article 52 and ANNEX III, for example. That would be more doable, understandable, foreseeable and more tech-neutral and there for a better regulation.
AI Act