German Digital Minister Karsten Wildberger (CDU) has alerted the public to the dangers posed by increasingly autonomous artificial intelligence (AI) systems, demanding that AI companies significantly improve the safety of their technologies. Speaking on Tuesday to RTL and ntv, Wildberger stated that the last two months have seen a worrying rise in incidents where AI agents have operated spontaneously and performed actions they were not meant to take. He stressed that these risks must be taken “very, very seriously,” adding that the world is not a testing ground for products; instead, they must be secure before being released.
Wildberger emphasized that the primary responsibility lies with the companies, which need to focus their efforts on bolstering system security. “For now, it’s up to the companies to ensure the products are safe,” he asserted. Concurrently, he highlighted concrete steps being taken in Germany, noting that the country is currently developing a dedicated AI Safety Institute in collaboration with leading researchers in the field.
Looking toward regulation, Wildberger underscored the necessity of international cooperation, naming the USA, China, and Europe as crucial partners. Since AI does not respect national borders, he argued that countries must jointly strategize on potential risks-such as cyberattacks by criminals-and potentially agree not to develop or make certain capabilities available. While advising Europe, he warned that remaining inactive constitutes a risk in itself, emphasizing the need for the continent to develop its own solutions.
