Written by zgiaonews• November 23, 2023• 2:51 pm• AI

### Empowering AI in Europe: A Group’s Call for Action

HomeAI### Empowering AI in Europe: A Group’s Call for Action

Representing the digital technology industry in Europe

There are still two key considerations to address as the AI Act nears its final stages of discussions. Firstly, the management of general-purpose artificial intelligence (GPAI) and foundation models, and secondly, the potential misalignment of current sectorial policies.

In Europe, only 8% of businesses are utilizing AI, significantly below the Commission’s target of 75% by 2030, with a mere 3% of EU-based AI unicorns in existence. The utilization of Artificial Intelligence by Western businesses and citizens in critical sectors such as natural tech, healthcare, manufacturing, and energy is vital for Europe’s competitiveness and economic resilience.

To position Europe as a global leader in the digital landscape, it is imperative for businesses to spearhead AI innovation, leveraging foundation models and GPAI. As representatives of the German online industry, we recognize the substantial potential in foundation models and the rise of innovative players, many of whom originate from Europe. It is essential not to stifle their growth or prematurely regulate them out of existence before they have the opportunity to scale.

The Commission’s data indicates that even the introduction of a single AI-enabled product into the market could result in compliance costs exceeding €300,000 for an SME with 50 employees under the AI Act. Efforts must be made to mitigate this burden and enable SMEs to integrate GPAI, foundational models, and other advanced AI technologies into their operations.

We endorse the current initiatives of Member States to confine the application of foundation models to accountability standards. We firmly advocate for the regulation’s targeted approach, concentrating on high-risk applications rather than specific technologies. It is not necessary for the AI Act to govern every emerging technology.

Moreover, given the stringent regulations already in place such as the Medical Devices Regulation, it is crucial to address and eliminate any overlaps and discrepancies to ensure clarity.

Our Recommendations:

Maintain the risk-based approach as the foundation of the AI Act. The regulatory framework should remain technology-agnostic, focusing on high-risk use cases supported by a broad coalition of businesses and civil society. It should not attempt to regulate all AI software indiscriminately as high risk.
Addressing regional regulatory inconsistencies, particularly in the medical sector, is paramount. To mitigate conflicting requirements and overlaps and avoid disruptions to well-established regional frameworks, alignment with the EU’s comprehensive product safety legislation is essential.
Emphasize detailed disclosure, collaboration, and support for compliance throughout the value chain to regulate GPAI and foundation models effectively. Companies should be able to delineate engagement practices and allocate responsibilities under the AI Act. The concept of “readiness” or “high-impact” foundation models lacks quantifiability and does not guarantee future success.
The robust copyright protection and enforcement framework of the EU already contains provisions that can be leveraged to address AI-related issues, such as exemptions for text and data mining.

Joint Letter Signatories: