(de-news.net) – Germany has argued that Europe must bolster its technological sovereignty while addressing new security threats, rejecting calls to halt AI development. Public and corporate concerns about democracy, misinformation, employment, and loss of control are growing as AI adoption speeds up in German business, underscoring the conflict between effective oversight and technological competitiveness.
The German government has rejected the proposal of leading American artificial intelligence firms for a slowdown in research and development, arguing that such an approach would not be feasible for Europe. A spokesperson for the Federal Digital Ministry said the Federal Government’s objective was to strengthen digital sovereignty, a goal that also required Europe to continue advancing its own AI capabilities rather than deliberately restraining them. At the same time, Berlin is closely monitoring developments worldwide and taking warnings from leading AI developers seriously. The concern is particularly acute as increasingly powerful AI models may become capable of escaping controlled testing environments and independently accessing other systems, potentially creating an entirely new category of security threat.
Germany is responding in part by expanding its national AI Security Institute, which is intended to enable the early identification and mitigation of security-related risks associated with the technology. The government is also maintaining close exchanges with comparable institutions in partner countries, reflecting an effort to address risks through international cooperation as well as domestic capabilities. CDU Digital Minister Karsten Wildberger has recently advocated global oversight of AI, and Berlin intends to continue pursuing the international dialogue surrounding the governance and security of increasingly capable systems.
The debate is unfolding amid growing concern about Europe’s potential marginalization in the global competition over artificial intelligence. Prominent economists and politicians have warned that Europe could fall permanently behind without decisive action. A group led by former EU Commissioner Margrethe Vestager and Monika Schnitzer, chair of the German Council of Economic Experts, has proposed a strategy requiring investments of 1.3 trillion euros and up to 45 gigawatts of additional energy capacity for data centers by 2030. The scale of those requirements illustrates the infrastructure demands associated with maintaining technological competitiveness. Vestager has argued that Europe must avoid becoming a playing field for foreign powers.
At the same time, concern about the potential dangers of AI is becoming more pronounced. Artificial-intelligence expert Thilo Hagendorff has argued that warnings about a possible extinction of humanity should not be dismissed as exaggerated, while emphasizing that such predictions remain subject to substantial uncertainty. Clemens Fuest, president of the Ifo Institute, regards demands for slower AI development as justified in principle but doubts that a successful international agreement could realistically be implemented. Research is inherently difficult to monitor, he argues, while China is unlikely to participate in an arrangement designed to constrain the pace of technological development.
Fuest has characterized the competition for technological leadership as extending beyond rivalry among leading American companies to the broader strategic contest between the United States and China. Possession of the most advanced AI systems therefore has implications not only for economic performance but also for national security. Against that background, he argues that simply attempting to slow research would not resolve the underlying problem. Instead, technological development would need to proceed while simultaneously placing greater emphasis on reducing the security risks associated with increasingly capable systems.
German companies embrace AI as concerns over jobs and survival grow
Meanwhile, AI adoption has expanded sharply across the German economy. According to a Bitkom survey, 57 percent of German businesses now use AI, almost three times the proportion recorded two years earlier. Another 38 percent are either planning or discussing its introduction. AI has consequently moved beyond a relatively small group of technological pioneers and become increasingly established throughout the business sector. The speed of this diffusion, however, has been accompanied by considerable uncertainty about both its implications and the extent to which individual companies can adapt.
Around one-third of businesses now fear that AI could threaten their own survival, compared with 23 percent the previous year. At the same time, 91 percent identify AI as the most important technology of the future, while 80 percent primarily regard it as an opportunity for their companies. Employee training has also expanded substantially: 70 percent of businesses report that they are training their staff to use AI, compared with 43 percent the previous year. The figures point to a combination of strong expectations about the technology’s economic importance and growing concern about the consequences of failing to adapt.
Public concerns are similarly pronounced, particularly regarding democracy and the reliability of information. According to a Forsa survey conducted for RTL and ntv, 81 percent of Germans consider the risk that AI could be misused in ways that threaten democracy to be high. Concern is especially pronounced among people over 45, with 84 percent viewing the danger as high, compared with 74 percent among those aged 18 to 44. Women also express somewhat greater concern than men, indicating that perceptions of AI-related democratic risks vary across demographic groups.
The authenticity of digital content represents an even greater source of public anxiety. Some 89 percent of respondents are seriously concerned that they may in the future be unable to reliably distinguish genuine images, videos, or news reports from material generated or manipulated by artificial intelligence. Among respondents aged 18 to 29, the proportion rises to 97 percent. The findings suggest that uncertainty over the reliability of digitally mediated information has become one of the most immediate public concerns associated with the expanding use of AI.
Employment represents another significant area of uncertainty. One-fifth of working Germans are highly or very highly concerned that AI could assume a substantial portion of their current professional responsibilities within the next five years. The concern is more widespread among women, at 25 percent, than among men, at 15 percent. It is also more pronounced among workers aged 45 and older than among those between 18 and 44, suggesting that perceptions of occupational disruption differ both by gender and age.
Business adoption, nevertheless, continues to accelerate. Bitkom reported that 57 percent of German companies now use AI, compared with 36 percent a year earlier and only 20 percent two years earlier. The share of companies for which AI is not an issue has meanwhile fallen sharply. Only 4 percent now say the technology is irrelevant to their business, down from 17 percent the previous year and 41 percent two years earlier. AI is therefore becoming increasingly difficult for businesses to treat as a peripheral technological development.
As AI grows more autonomous, experts warn of difficult-to-control systems
Despite this widespread adoption, most companies remain at an early stage of exploiting the technology’s capabilities. None of the companies using AI said that it was making full use of its potential, while only 3 percent believed that they were exploiting that potential to a relatively large extent. By contrast, 28 percent said they were making little use of AI’s possibilities, while 59 percent said they were making virtually no use of them. The figures indicate that adoption itself does not necessarily translate into mature or extensive use of the technology.
Fundamental skepticism about AI has nevertheless declined considerably. Ninety-one percent of companies now consider AI the most important future technology, compared with 81 percent last year and 73 percent in 2024. Only 8 percent regard AI as a hype that is being substantially overestimated, a marked decline from 26 percent in 2024. From the perspective of their own businesses, 80 percent primarily see AI as an opportunity, while 18 percent regard it more as a risk. No company surveyed considered AI overwhelmingly risky, and none believed that the technology would have no impact on its business.
Yet the expansion of AI is accompanied by substantial uncertainty. Sixty-one percent of companies believe that businesses without AI use have no future, compared with 51 percent in 2025, while 56 percent believe Germany has already fallen behind in AI. Thirty-four percent fear that the technology could threaten their company’s survival, up from 23 percent the previous year, and 31 percent expect AI to alter their business model, compared with 24 percent in 2025. Among companies already using AI, one-third, or 33 percent, acknowledge that fear of falling behind is a primary reason for their adoption of the technology.
Hagendorff has identified potential scenarios in which control over AI systems could be lost, particularly where multiple AI agents operate collectively as a network, have access to the Internet, and are no longer subject to effective monitoring. He has also warned that oversight could become more difficult as it becomes harder to interpret the internal processes through which AI systems reach their conclusions or pursue their objectives. The University of Stuttgart researcher argues that AI systems must therefore become not only more capable but also better able to incorporate moral considerations comparable to those used by humans. Although AI alignment—the effort to orient systems toward human goals and ethical rules—has achieved substantial progress, he sees particular challenges in multi-agent systems capable of establishing and pursuing their own objectives.
Several prominent American AI developers have recently called for development to be slowed. More than 1,000 employees of major companies had previously signed a petition urging the U.S. government to deliberately reduce the pace of automated AI development. US President Donald Trump has indicated that legal guardrails could be possible while emphasizing that the United States must remain the leading nation in AI. The debate thus reflects a tension between demands for greater caution and the strategic importance attached to maintaining technological leadership.
OpenAI and Anthropic have also raised concerns about the trajectory of AI development. Anthropic researchers Jacob Coxon and Evan Hubinger have warned that a self-improving AI system could potentially pose an existential threat to humanity. Jakub Pachocki, OpenAI’s head of research, has likewise urged extreme caution, arguing that machine intelligence is beginning to surpass human intelligence. He has also emphasized that AI may eventually become necessary to defend against attacks carried out by other AI systems, adding a further dimension to the emerging security debate.
Some observers, however, interpret such warnings partly in light of the commercial interests of AI companies. By emphasizing extreme future scenarios, providers may simultaneously draw attention to the capabilities and strategic importance of their own systems. The warnings nevertheless come amid concrete incidents that have intensified the security debate. Several cases have emerged in which AI systems reportedly penetrated external IT infrastructure and deliberately deceived people in order to accomplish assigned objectives. Such incidents have added practical evidence to broader concerns about the security implications of increasingly autonomous AI and have reinforced the debate over how development can proceed without allowing technological capability to outpace effective oversight.
