At TechBBQ, Europe’s AI conversations kept coming back to: Who’s actually in control?

At TechBBQ in Copenhagen, the discussions among investors, founders, and operators from across Europe centered on a pressing issue: who truly controls the AI revolution. The annual Nordic conference, themed “Emerging from Agency,” saw recurring conversations about the implications of AI sovereignty, particularly following the recent unavailability of Anthropic’s AI models Mythos and Fable outside Europe. These developments sparked intense debate about whether Europe should develop its own AI infrastructure or continue relying on American and Chinese counterparts. The absence of Mythos and Fable from European markets highlighted the broader concerns about dependency on foreign AI giants. One startup executive described the disruption caused by the models’ unavailability, while others remained pragmatic, acknowledging the risks but noting that current operations remain functional. The situation underscored a growing sentiment that Europe must assert greater control over the AI landscape to avoid potential long-term repercussions. Ellen de Brever, an angel investor and head of partnerships at the Novo Nordisk Foundation Cellerator, emphasized that the conference’s theme of agency resonated deeply. She noted that the discourse had evolved beyond merely exploring AI capabilities to questioning what aspects of decision-making Europeans are willing to relinquish. De Brever observed that the conversations were less about enhancing machine intelligence and more about determining who retains authority over AI’s deployment. During the event, a panel featuring Signal President Meredith Whittaker addressed the intersection of privacy and AI. Whittaker criticized the integration of AI assistants into operating systems, citing concerns about data collection and user consent. She argued that AI labs often obscure the ethical consequences of extensive data acquisition, urging a renewed emphasis on privacy as a critical component of AI governance. Emad Mostaque, co-founder of Stability AI and Intelligent Internet, joined a discussion on the societal impacts of an agentic workforce. He warned of the concentration of power in the hands of a few AI labs, suggesting that control over AI equates to control over national policies. His remarks prompted reflection from attendees, including Mia Negru of Life With Artificials, who mused on the broader implications of autonomous AI systems. She speculated that such advancements could necessitate a reevaluation of fundamental concepts like ownership, labor, and democratic participation. Beyond formal sessions, the conference fostered informal exchanges, with tech leaders and investors engaging in relaxed discussions at happy hours and rooftop gatherings. Topics ranged from Denmark’s rising tech hubs to collaborations between Nordic and African innovators. Notable participants included Lovable, Nvidia, OpenAI, AWS, and HSBC, highlighting the diverse interests converging on the event. Meanwhile, Anthropic made strides in advancing AI research with a new paper outlining a system capable of improving AI alignment through automation. Led by fellow Chen Yueh-Han, the project demonstrated that AI-driven researchers could enhance model performance across multiple alignment benchmarks without compromising overall effectiveness. The findings suggest a path toward recursive self-improvement, potentially rendering human researchers redundant in the near future. The paper also compared the efficiency of automated systems to human-led research, noting that automated alignment researchers outperformed human experts within six hours. Cost comparisons further underscored the economic advantages of AI-driven solutions, with automated systems costing approximately $4 per hour versus $150 for human researchers. Despite these benefits, the study acknowledged limitations, emphasizing the need for robust benchmarks and continuous refinement of the underlying literature. In parallel, the Pentagon expanded its use of AI tools by introducing military versions of ChatGPT and Grok. These customized platforms aim to support operational tasks while avoiding the data collection practices typical of consumer AI applications. The initiative reflects a strategic effort to harness AI for national security purposes, with additional partnerships forming with companies like Amazon Web Services, Microsoft, and Nvidia. OpenAI also announced plans to integrate ChatGPT into healthcare systems, partnering with Epic to streamline clinical workflows. The tool allows clinicians to access patient data, generate summaries, and prepare for appointments, all while adhering to strict read-only protocols. Additionally, OpenAI introduced a Healthcare Public Data plug-in, enabling access to medical databases and aiding in clinical decision-making. As major AI firms vie for dominance in government, healthcare, and economic sectors, the race for control over AI continues to intensify. The ongoing dialogue at TechBBQ underscores the complexity of navigating this rapidly evolving landscape, where questions of sovereignty, ethics, and responsibility remain central to the conversation.