Three researchers at the Austrian Academy of Sciences call for stronger institutional responsibility for AI. In an October 9, 2026 original interview, they contrast current power structures and verifiable oversight with a focus on catastrophic future scenarios.
Social scientist Astrid Mager asks who can conduct independent reviews, what expertise evaluators have and what access they receive. Responsibility should not be placed solely on individual staff assigned human oversight tasks, she argues.
Information systems researcher Stefan Strauß recommends effective supervisory authorities and minimum standards. Institute director Michael Nentwich favors European capacity and alternatives to a few dominant providers.
The researchers offer assessments and proposals. The interview neither measures technical loss of control nor future extinction risks, and it announces no new binding rulebook.