European Union lawmakers are strongly opposed to potential changes to Bloc’s AI regulations, warning that such moves could exempt major tech companies such as Openai and Google from key compliance obligations.
According to the Financial Times, the European Commission is considering making certain provisions of AI Act voluntary rather than mandatory. These regulations are designed to ensure that advanced AI models do not generate harmful or misleading content and are not misused due to election interference. The proposed changes follow important lobbying from American tech giants and politicians, including former US President Donald Trump.
Several members of the European Parliament (MEPS) involved in drafting the AI Act have expressed concerns about these revisions in a letter to the committee’s digital chief, Henna Virkkunen, according to the Financial Times. The MEP argues that weakening regulations could lead to increased risks such as foreign interference, election manipulation, discrimination, and spreading illegal content. They also warn that such changes could have a serious impact on European economies and democracy.
The AI Act, which categorizes AI technologies based on risk levels, requires strict reporting requirements for high-risk applications, such as those currently used in healthcare and public transport. Additionally, powerful AI models must comply with additional transparency obligations regarding training data and methodology.
Related: As US trade disputes grow, the antitrust war on big technology in the EU gets heated
At the heart of the current discussion is a new “Composition Code of Practice” aimed at guiding AI companies to implement AI laws. The code is developed by a panel of experts, including Turing Award winner Joshua Bengio, and is scheduled to be adopted by the European Commission in May. However, finding a balance between strict enforcement and voluntary industry participation is a challenge.
The EU is facing intense lobbying from the US over AI regulations. Joel Caplan, head of global affairs for Meta, recently warned that the proposed code of practice could impose “an in-feasible and technically unfeasible requirements.” Meta also argues that due to EU privacy rules, it is not possible to deploy modern AI models within the block. Other companies, including Google, as well as European companies such as Spotify and Ericsson, have also criticised the regulatory approach.
Despite these pressures, Virkkunen reaffirmed the committee’s commitment to maintaining a fair and democratic digital landscape in Europe. At a recent event, she emphasized that codes of practice should provide clear guidance to industry and stakeholders rather than introducing additional hurdles.
As the debate continues, the EU faces important decisions. It’s either maintaining its stance as a global leader in AI regulation or yielding the industry’s demands for deregulation.
Source: Financial Times
