In this world of visual digital media moving and being shared at the speed of sound, connectivity has definitely been improved, but it comes with some concerns. Visual media manipulation technologies have also evolved, giving amateurs and experts alike the tools to realistically manipulate media for potentially nefarious ends. This interesting information came to us from Government CIO in their article, “DARPA Tackles Deepfakes With AI.”

The Media Forensics (MediFor) program out of DARPA’s Information Innovation Office (I2O) is trying to develop automated artificial intelligence (AI) technologies to assess visual media manipulation at scale in an end-to-end platform. While manual media forensics can be time-consuming and requires expertise to identify manipulation, the MediFor platform would apply media forensics techniques automatically.

To undertake automatic media manipulation detection, the program has established a framework of three tiers of information, including digital integrity, physical integrity and semantic integrity. Semantic integrity constraints may describe properties of and relationships between data objects and may also place limitations on permissible data base operations.

Melody K. Smith

Sponsored by Data Harmony, a unit of Access Innovations, the world leader in indexing and making content findable.