German Bundestag, Plenarprotokoll 20/179, pp. 23264-23265.
Frau President. Right honorable ladies and gentlemen. Valued colleagues.
After a good nine months, we here in plenary session again deliberate on the Union motion for a strengthening of Artificial Intelligence as a key technology for Germany. After nine months, it certainly can be expected that a new thought or a new idea will be born. Yet unfortunately, despite a thorough, orderly description of the problem: A zero return in your motion.
Ladies and gentlemen, what becomes clear here, and what is it actually about? It’s about no more and yet no less than Germany’s role in the AI and data market. So as to bring it to three points: First, start-ups and KMU [Kleine und Mittlere Unternehmen, small and mid-sized business] require free development possibilities without sprawling bureaucratic impediments. Instead is required, second, more innovation and knowledge transfer, and with that, third, a durable business model with scaling perspectives.
The 2024 Federal Research and Innovation Report, which here yesterday in plenary session was already presented, lays the finger in the wound. In regards the themes of data and AI, the Federal government neglects the necessary computing infrastructure. It does not do to rebuke the without question splendid achievements of German research institutions in the AI area.
We in Germany have no recognition or knowledge problem when we speak of AI. Yet, right honorable ladies and gentlemen, what is lacking is a secure, stable computing infrastructure so as to train the algorithms of quite promising start-ups and KMU. For that, these so far need to use the services of large hyperscalers, with the real risks that their training data does not remain exclusively with them. The local supercomputers, for instance in Jülich, Garching or Stuttgart, are primarily computers for science which are only at limited disposal for young business customers.
There continues to be lacking in Germany a securely financed AI infrastructure. The AI-Federal Association LEAM [large European AI models] initiative figures the construction costs of one such at around 300 million euros. The Federal government, which so gladly styles itself as progressive coalition, has so far kept from view this practical stimulus. To name this refusal is the duty of a constructive opposition.
The coalition, continues to appear to be, ja, of good cheer to decide in the coming weeks in cabinet the draft budget for 2025. For the digital, data and AI policy, the budget is allowed to become a tragedy: Reductions here, eliminations there, minimal investments. How, with that, valued ladies and gentlemen, shall be strengthened Germany’s role in regards the key technology of AI, remains a secret.
Otto Fricke
(FDP): The AfD wants to expend more money! Always good!
And the once loudly announced digital budget was in the meantime well buried. Or will the government still apply itself, valued Herr colleague Broadcasting Kaiser?
As was already adhered to last autumn in this place: Your motion, valued colleagues of the Union, aims in the right direction. It nevertheless is lacking in exactitude and consistency. We therefore will abstain. Our offer, after the coming Bundestag election in a coalition of the reasonable to deal with the AI theme with corresponding priority, stands as before.
Many thanks.
[trans: tem]
