With the increasing proliferation of AI-generated images and videos, the risk of targeted historical falsification also grows. Historical recordings and Holocaust documents are particularly affected, being increasingly misused as the basis for deceptively real deepfakes. A team of students from the Ludwig Maximilians University Munich (LMU) now aims to address this with a new technological solution.

The "Global Authentic Memory Initiative" (GAMI) is developing a system to ensure the authenticity of digitized historical documents in the long term. The goal is to support archives, memorials, and documentation centers in protecting their digital holdings from manipulation and later discrediting.

At the center are three publicly verifiable properties: the integrity, existence, and institutional origin of digital files. For this, GAMI generates cryptographic fingerprints of so-called hash values, which immediately reveal changes to files. Additionally, these proofs are anchored in the Bitcoin blockchain, which can prove that a document existed before modern AI deepfakes were technically possible.

Another component is the cryptographic signature by the respective institution. This allows verification of which archive or memorial a file comes from independently, without requiring a central controlling authority.

System initiated by LMU students

The initiative was founded at the end of 2025 by LMU students Pablo Toussaint, Florimon Poisson, and Rosa Bauer. The motivation is the concern that even genuine historical recordings could be discredited as alleged AI forgeries in the future. Particularly in the context of Holocaust remembrance, the initiators see an urgent need for action.

GAMI is already working on collaborations with institutions such as the Federal Archives, the NS Documentation Center in Munich, and concentration camp memorials, as well as international partners like Yad Vashem in Jerusalem. The project is also supported by scientists from LMU, TUM, and Princeton University.

The initiative exemplifies how technological innovations can be used to protect digital memory culture and historical truth in the age of generative AI in the long term.