Truth Engine
Exploring AI, Ethics & Information Integrity
In a world where synthetic media—deepfakes, AI-generated voices, and manipulated images—are becoming alarmingly convincing, Gretchen AI is pioneering a new kind of digital defense. Our intelligent agents are designed not just to detect manipulation, but to understand context.
Gretchen AI Wins Cybersecurity Pitch at VivaTech 2025
Daniel Röder shared our vision for using AI against disinformation — and Gretchen AI won first place in the cybersecurity pitch!
Advancing to Phase 2 of the SPRIND Funken Challenge
Gretchen AI selected among top teams developing real-time deepfake detection technology
by Yassine El Kheir
BiCrossMamba-ST:
Speech Deepfake Detection with Bidirectional Mamba Spectro-Temporal Cross-Attention.
AI can mimic voices from minutes of audio. BiCrossMamba-ST by Gretchen’s Team member Yassine El Kheir beats top deepfake detectors with fewer parameters and up to 67.7% higher accuracy.
Generalizable Audio Spoofing Detection using Non-Semantic Representations
Deepfake audio detectors often fail in real-world use. Arnab Das’ team used non-semantic audio features that capture voice traits and emotion but ignore speech content.
SPRIND Backs Gretchen AI in Deepfake Detection & Prevention Drive
SPRIND’s Deepfake Detection and Prevention program seeks quick, impactful ideas. Gretchen AI, a DFKI spin-off, and dpa are partnering to help newsrooms tackle AI challenges.
Gretchen AI featured in #ai_berlin blog!
We are thrilled to be featured in the #ai_berlin blog! Fighting deepfakes and disinformation requires precision, efficiency, and the right AI tools. At Gretchen AI, we develop intelligent agents that automate fact-checking where it matters most. Thanks to hashtag #ai_berlin for this special spotlight on our work!
From the lab to the street
Gretchen AI founders Daniel Röder, Arnab Das, and Dr. Tim Polzehl bring their many years of expertise as researchers at the DFKI Lab Berlin to the company. As part of news-polygraph, Germany’s largest research project on media content verification, they collaborate with renowned public broadcasters as project partners.
Gretchen AI featured in #ai_berlin blog!
We are thrilled to be featured in the blog! Fighting deepfakes and disinformation requires precision, efficiency, and the right AI tools. At Gretchen AI, we develop intelligent agents that automate fact-checking where it matters most. Thanks to hashtag for this special spotlight on our work!
SPRIND Backs Gretchen AI in Deepfake Detection & Prevention Drive
The SPRIND – Federal Agency for Disruptive Innovations – is now taking a fresh approach in every respect with the Deepfake Detection and Prevention program. Startups are seeking the best and, above all, quickly implementable ideas. Among them is Gretchen AI, a spin-off of the German Research Center for Artificial Intelligence (DFKI) . And we at dpa are involved as a partner. Together we want to develop solutions that also help newsrooms overcome the challenges posed by AI.
From the lab to the street
Gretchen AI founders Daniel Röder, Arnab Das, and Dr. Tim Polzehl bring their many years of expertise as researchers at the DFKI Lab Berlin to the company. As part of Newspolygraph, Germany’s largest research project on media content verification, they collaborate with renowned public broadcasters as project partners.
Strategic Alliances Powering Gretchen AI’s Mission
At Gretchen AI, we know the fight against disinformation and digital manipulation can’t be won alone. That’s why we’ve partnered with leading voices in AI research, journalism, and public innovation to build a trusted ecosystem rooted in accuracy, transparency, and integrity.
Germany’s largest AI research center, with 100+ spin-offs and 500+ AI industry and research projects.
Germany’s federal agency for disruptive innovation, supports us financially and strategically to develop solutions that anticipate and shape the future of trustworthy information.
Germany’s largest news agency, gives Gretchen AI access to its vast Picture Alliance database, enabling fast, large-scale visual content verification to uphold media integrity in real time.