Berlin gets it right: Putting truth first
Berlin, May 27, 2026 – When Gretchen famously asks Faust, “Now tell me, how do you feel about religion?”, she puts him on the spot, pushing him to come clean about what he really believes. Since then, the “Gretchen question” has become shorthand in German for a crucial, no-nonsense question that lays bare one’s real position or convictions – or, in the case of Berlin startup Gretchen AI, the truth itself. The young company, a spin-off from the German Research Center for Artificial Intelligence, develops technologies that detect manipulation in images, videos, and audio files while also analyzing the context of digital information. Its goal is to uncover disinformation faster and more reliably – while reinforcing something increasingly in short supply: trust.
Founded in 2024 by AI expert Dr. Tim Polzehl, DevOps specialist Arnab Das, and AI agent and RAG expert Daniel Röder, the company brings together a team of former DFKI researchers. Their shared mission is to simplify the verification of digital information and counter disinformation using advanced AI tools. The challenge is widely felt. Whether in articles, videos, or images, distinguishing fact from fake is becoming ever more difficult – not just for the general public, but also for those who curate and verify information: journalists, editors, public authorities, and institutions alike. Drawing on years of research, Gretchen AI has developed a platform that combines AI agents, deepfake detection, and context-based analysis. With a single query – “Gretchen question” – users can run complex checks, such as verifying the authenticity of an image or video. The technology goes beyond traditional deepfake detection. To identify the original context of a piece of media, forensic analysis is paired with automated online research. The results are backed by sources – a critical advantage, particularly for journalists and organizations engaged in fact-checking. The platform also includes features for knowledge management, AI-powered search, and content creation, making it easy to integrate into professional editorial workflows.
Importantly, the system does not simply label content as “true” or “false.” Instead, it provides context – often essential for understanding and assessing complex information. The team led by Tim Polzehl believes that fact-checking should become a routine part of media consumption – but it must also become significantly easier to carry out. Gretchen AI’s automated solution saves both time and resources. As part of Newspolygraph, a German research initiative focused on AI-based detection of disinformation and deepfakes, the company is already working with public broadcasters. In addition, the team is collaborating with the Deutsche Presse-Agentur to build a reference database of verified facts.
The company’s innovative approach has not gone unnoticed. In 2025, Gretchen AI received the Deepfake Detection Challenge Award from Germany’s Federal Agency for Breakthrough Innovation and was nominated for the 2025 Berlin Brandenburg Innovation Award. This May 2026, they also won the Deep Tech Award in the “Artificial Intelligence” category — an award that recognizes outstanding companies whose innovations are shaping the future.
Just one year after its founding, the startup appears well positioned for further – verified – milestones ahead.