AI, cloud, chips: Heads of state put pressure on digital sovereignty

Against the backdrop of "increasing economic and security policy challenges", the federal states want to work more closely with the federal government to secure the technological sovereignty that has often been invoked for years and to promote artificial intelligence (AI) as a "central cross-sectional and future technology" in Germany and Europe. The heads of government of the federal states passed a resolution to this effect at their conference in Berlin on Wednesday. In it, they emphasize that " digital sovereignty is a fundamental prerequisite for technological and economic competitiveness and security".

In the paper, the heads of state criticize "the dependence on non-European technology providers in key areas" such as AI, cloud infrastructures and high-performance chips. This unfortunate situation not only jeopardizes innovation, "but also the autonomy of state and economic actors". To meet this challenge, they believe that "greater support for open and interoperable technologies" is necessary, especially in AI development.

According to the minister presidents, open source models and open source software and hardware offer an option to "create competitive, secure and transparent alternatives to proprietary systems". At the same time, it is necessary to "establish binding open standards for AI models, interfaces and data formats in order to guarantee sustainable digital sovereignty".

Proprietary technology as a security risk

Proprietary systems without full access to the source code could pose a "significant risk to the IT security of critical infrastructures and security-relevant AI applications", the heads of government state. Models for this area must therefore meet the necessary security standards and be able to be independently checked for "security risks, distortions and manipulation possibilities".

The aim must be to advance from user to creator and from taker to giver of technologies, they continue. It is crucial to translate research results into concrete products and applications and to create a framework for "own cross-domain AI models and cyber-secure, easy-to-use systems and applications". As part of the national AI strategy, the German government should therefore create opportunities for the use of anonymized and pseudonymized data and give greater consideration to decentralized system approaches. It is important to avoid transferring sensitive data to central servers.

At the same time, the heads of the federal states are asking the federal government to "specifically improve the framework conditions for the construction and operation of data centers and other AI infrastructure". A long-term strategy is required not only to enable the establishment of such data centers, but also to plan targeted follow-up investments – in digital parks or concepts for the use of waste heat –. European providers must be given appropriate consideration and targeted support.

Digital economy praises the national initiative

"Europe must also strengthen its technological independence when it comes to high-performance chips for AI applications," the paper states. To drive forward German research in the field of new, energy-efficient chips – such as neuromorphic hardware or photonic semiconductors – in a targeted manner, both the market maturity of this new generation should be accelerated and the establishment of corresponding chip factories should be intensively promoted.

However, the use of AI in social media and the associated disinformation also represents a growing challenge for democracy, the heads of government state. Comprehensive measures to promote digital and media literacy are therefore important. However, excessive regulation could prevent innovation with the help of AI. The AI regulation should therefore be implemented as unbureaucratically as possible. However, the technology requires "solutions for the protection of intellectual property".

The IT association Bitkom generally welcomed the decision. According to Bitkom, the federal states should place AI at the center of their funding activities and promote cooperation between universities, tech start-ups and established companies more strongly. The establishment of more English-language AI degree courses is also necessary. The new federal government should not only continue the federal government's AI strategy, but also release more funds for it and, above all, make them available in an uncomplicated manner.

(wpl)