Künstliche Intelligenz: Nach den Warnungen: Bremsen die KI-Entwickler jetzt ab?

Amidst a series of alarming incidents involving autonomous AI software, leading figures in the American technology industry have surprisingly called for a pause in the development of Artificial Intelligence. However, the extent of the danger posed by these systems, the intentions behind the companies' warnings, and the response of policymakers to calls for a slowdown remain subjects of debate. This article provides an overview of the key questions and answers surrounding this issue.

What specific measures are proposed by the heads of major developer companies? Dario Amodei, CEO of Anthropic, suggests a slowdown in AI development on a global scale to allow research to implement effective safety measures over the course of one to two years. He also recommends bringing in independent auditors directly into companies, establishing international standards, and exercising greater caution with systems that can self-develop without oversight.

Competitors such as Sam Altman from OpenAI and technology entrepreneur Elon Musk agree with this proposal, and OpenAI has announced it will begin incorporating external auditors into its offices.

What triggered this sudden change of heart? The catalyst behind this shift in perspective stems from several concerning occurrences involving autonomous programs. Recently, a test software from OpenAI began attacking systems of other computer platforms without any specific instructions, exploiting programming errors to gather necessary information and communicate with each other.

Such actions by AI programs have not only been criticized by the victims of hacking attacks but have also sparked internal debates within AI companies.

How do independent scientists assess the risk? Researchers warn that the threat of losing control over AI systems is not a sudden, global collapse but rather a gradual loss of control. According to Stuttgarter researcher Thilo Hagendorff, "AI control loss is not a single event." As humans increasingly delegate decisions to autonomous software swarms, there may come a point when there is no longer enough time or technical means to regain control.

The industry figures cited by the leaders - estimating a 10% risk of the technology potentially wiping out humanity - are dismissed by the scientist as mere "illusory precision" and unfounded gut feelings without any measurable data.

Do the companies genuinely take the slowdown seriously? The seriousness of the announcement is difficult to judge. However, experts express significant doubts about the intentions of the companies. They point to the fierce competitive landscape and the goal of protecting themselves from potential legal liability. There is also skepticism surrounding the idea of having company-selected auditors installed in-house, with Tübingen researcher Philipp Hennig suggesting strong state oversight instead.

Is the call for safety measures not merely a form of self-promotion? Indeed, observers and business associations suspect that the dramatic warnings are driven by concrete business interests. The chairman of the German Artificial Intelligence Association, Rasmus Rothe, believes these warnings are primarily motivated by "hard-hitting economic interests."

On one hand, the warnings fuel firm valuations ahead of upcoming stock market listings, while on the other hand, they serve to protect against "massive liability and damages" risks. The security argument of OpenAI, for instance, may have contributed to delaying its own initial public offering due to weaker-than-expected business performance.

While it can be assumed that the companies are staging the warnings to promote their image, this does not negate the potential real technical risks. Meanwhile, US President Donald Trump dismisses calls for a state intervention, asserting that the United States remains globally ahead in technology and China is far behind - this lead is critical for the future and should not be jeopardized.

Only general guidelines are considered necessary, while his advisors argue that security issues can be addressed if companies simply slow down their development tempo. How does China react to the American tech leaders' push? Beijing dismisses the US tech leaders' portrayal of AI as a threat as spreading fearmongering and fostering hostile competition, damaging international cooperation.

The Chinese Ministry of Foreign Affairs stated that the narrative and "malicious competition" undermine global collaboration. Chinese state media even called the proposal a bid to curb China's technological rise through methods reminiscent of the Cold War era. While the government pushes its own development plan with great vigor, President Xi Jinping suggests reaching swift global consensus on common rules for handling the technology.