Вештачка интелигенција

A New Perspective on Artificial Intelligence in the European Union

Summary

In this weekly review, we bring you the latest news on artificial intelligence in the European Union, offering a fresh perspective on the topic and highlighting key facts and updates from the sector. Firstly, it is important to note that […]

A New Perspective on Artificial Intelligence in the European Union

In this weekly review, we bring you the latest news on artificial intelligence in the European Union, offering a fresh perspective on the topic and highlighting key facts and updates from the sector.

Firstly, it is important to note that artificial intelligence is at the center of negotiations surrounding the AI Act, which is currently in the process of being drafted. One of the key topics of discussion is foundational models of artificial intelligence.

Instead of quoting the European Parliament, foundational models can be described as models trained on a large amount of data and capable of executing a wide range of tasks.

During the past week, the Spanish Presidency of the EU Council presented its proposals for foundational models for the first time, employing a category-based approach. The proposed framework includes a strong reference to the Copyright Directive, as well as the protection of copyright owners’ rights.

In response to these proposals, co-rapporteurs of the European Parliament put forth the first set of criteria for classifying foundational models as “highly impactful.” They have also suggested additional criteria for assessing systemic risks, shifting copyright provisions to an article dedicated to generative artificial intelligence. Furthermore, they highlighted the importance of clear responsibilities throughout the entire value chain.

Changes are also beginning to occur in the field of law enforcement. An internal document from the European Parliament proposes lifting the outright ban on real-time biometric recognition in favor of other types of prohibitions, such as categorizing biometric data and emotion recognition. However, it remains uncertain whether this change will be acceptable to left-leaning politicians, as Brando Benifei stated for Euractiv that this is just one of the possibilities and that there will be more proposals.

As part of the negotiations, the Spanish Presidency has also introduced a version of provisions related to AI governance, including the establishment of a scientific panel. However, the question of whether service providers themselves or external experts will be responsible for verifying the security of AI systems remains open.

Don’t miss: The Cyber Resilience Act is expected to receive significant political support by the end of November. Negotiators have already agreed on the support to be provided during this period, with the Parliament introducing the principle of proportionality and advocating for a minimum five-year support period. The only remaining question is to whom the first report should be submitted.

Furthermore, the OECD has officially updated its definition of artificial intelligence to assist in the development of new EU AI legislation. The definition has been adapted to market and technological changes in recent years, with a focus on AI systems supported by machine learning.

Germany also aims to enhance the development of artificial intelligence at the national and European levels, as outlined in a new action plan presented by the Ministry of Education and Research. The plan aims to propel the EU to be on par with the already dominant US and China in the field of AI. Key measures include strengthening the entire AI value chain, fostering collaboration with education and research, and a planned investment of 1.6 billion euros in AI by the end of the current government’s term.

In addition to negotiations and action plans, there have been numerous key terms related to artificial intelligence discussed in recent times. The EU and the US have released 65 key AI terms that are important for understanding risk-based approaches and definitions of AI. Their activities in standardization have also been mapped to identify common interests in AI-related standards.

The G7 nations held a meeting focused on competition in the AI sector. Competition authorities have adopted a declaration on competition in the digital sector, with a particular emphasis on AI. It is recognized that competition authorities worldwide are equipped with AI tools that can be used for proper oversight of the development of this sector.

Through these key news and trends, it is clear that the European Union is taking the issue of artificial intelligence seriously. Negotiations on the AI Act are intense, and member states and institutions are focused on creating legal regulations that will enable safe and ethical use of AI technologies. All these efforts aim to protect the rights and freedoms of citizens while promoting innovation and competitiveness in the digital age.

FAQ