Skip to content

Critique on the Proposed Artificial Intelligence Law by the European Commission

EU concludes consultation on the proposed Artificial Intelligence Act, set to proceed through Parliament and Council of Ministers. The Center for Data Innovation offers preliminary feedback on the Commission's strategies, aligning with their views.

Assessment of the Artificial Intelligence Regulation Proposed by the European Commission
Assessment of the Artificial Intelligence Regulation Proposed by the European Commission

Critique on the Proposed Artificial Intelligence Law by the European Commission

The Center for Data Innovation has submitted a detailed filing with suggestions for the proposed Artificial Intelligence Act, aiming to improve its clarity, feasibility, and alignment with the evolving regulatory framework for General-Purpose AI (GPAI) models.

The filing emphasises the need for a well-designed regulatory framework that encourages responsible adoption and use of AI, while addressing potential harms and fostering growth and innovation in the European digital economy.

One of the key recommendations made by the Center for Data Innovation is to provide clearer and proportionate operational measures tailored to the size and type of AI providers, including startups and Small and Medium-sized Enterprises (SMEs), to ensure feasible compliance without overburdening smaller entities.

Another suggestion is to introduce fixed documentation retention periods (e.g., 10 years) for AI model versions and require updates upon significant changes to improve transparency and traceability. The filing also proposes making certain non-sensitive information about GPAI models publicly available to foster trust and public understanding.

The Center for Data Innovation also advocates for strengthening obligations related to copyright compliance by requiring providers to reproduce and extract only lawfully accessible content, respect machine-readable rights reservations during web crawling, and implement technical safeguards to prevent output infringement. The filing further suggests establishing grievance mechanisms with clear points of contact for rights holders to enable effective redress against infringing outputs.

The filing also clarifies that adherence to voluntary codes related to these issues demonstrates alignment with the AI Act but does not substitute for full legal compliance with copyright laws.

Moreover, the filing underlines the importance of risk identification, assessment, and mitigation for systemic risks posed by GPAI models, supported by comprehensive evaluations including adversarial testing and cybersecurity measures.

These suggestions align with provisions found in the EU's Voluntary GPAI Code, which builds on and expands obligations in the AI Act, aiming to reflect a practical and balanced approach to regulation that considers technological and operational realities.

The proposed AI Act will be reviewed by the European Parliament and Council of Ministers. The Center for Data Innovation encourages a collaborative approach between the EU, Member States, and stakeholders in the development and implementation of the AI Act.

AI is a new frontier of digital technology with transformative impacts on the economy and society in the next decades. The filing argues that the AI Act should provide clear guidelines for the development, deployment, and use of AI, including ethical considerations. It also underlines the importance of creating conditions that allow citizens to take advantage of AI's opportunities for growth and progress.

The filing highlights the need for a risk-based approach in the regulation of AI, with appropriate distinctions made between high-risk and low-risk applications. It also recommends that the AI Act should be designed in a way that promotes transparency and accountability in AI systems.

In summary, the Center for Data Innovation likely recommends improving the AI Act by clarifying requirements, promoting transparency, ensuring copyright compliance, and including proportionate measures for different providers, as reflected in the evolving EU regulatory framework and voluntary guidelines for GPAI models. However, a direct detailed list from the Center for Data Innovation itself was not found in the search results. The filing also suggests that the AI Act should include provisions for cross-border data flows to ensure the free movement of data within the EU.

Read also:

Latest

LiDAR technology showcased by ZVISION SPAD at CES

CES ZVISION StaLiDAR Scanner Technology Unveiled

At the 2024 Consumer Electronics Show (CES), company revealed their newest SPAD LiDAR, the ZVISION EZ6, on the 2nd of January, making its debut in the public eye. Designed for mass production within Advanced Driver Assistance Systems (ADAS), the ZVISION EZ6 is a long-range and economical LiDAR...