The CEO of the German AI startup Aleph Alpha, Jonas Andrulis, has warned of the consequences of Europe’s lack of its own AI models in the face of the success of the Chinese AI, DeepSeek.
“We need to ask ourselves in Europe what it’s worth to train an EU-based LLM that is based on our rules and values, a model that wasn’t trained with Chinese government guidelines or US cultural norms” Andrulis told the Welt am Sonntag.
The most well-known Large Language Model is ChatGPT. The Chinese AI, DeepSeek R1, caused a stir earlier this week, as it appears to require significantly less computational power than rival models from the US. However, during tests, it was found to hold positions of the Chinese government, such as on Taiwan.
From the perspective of a liberal democracy, Andrulis views it critically when, for instance, young people engage with an AI that generates a world view only driven by the ideological or economic motivations of the US or China, he said. “I’m 100 percent sure, especially in education, that we need something of our own, but also in the cultural or social sphere” he emphasized.