Yesterday, we announced that Manifest has been awarded an Air Force Research Laboratory (AFRL) innovation contract to pursue the application of artificial intelligence bills of materials (AIBOMs) within the Air Force ecosystem. As far as we can tell, this represents the first dollars spent by the United States Government anywhere on AIBOM capabilities, and we could not be prouder for and more excited about this opportunity.
In this time of unprecedented AI adoption, it is absolutely critical that AI adopters inventory the model cards and lineages of their AI solutions. We will only have one opportunity to keep tabs on where each model is deployed, what it was trained on, and which software it uses. If we miss this opportunity, we as a society will find ourselves in exactly the same situation as we found ourselves on the traditional software side during the Log4shell incident: scrambling to inventory our technology after the fact.
Warfighters already know this. In August 2023, the Assistant Secretary of the Army Acquisitions, Logistics, and Technology, or ASA (ALT), put out an RFI seeking industry insight into AIBOM capabilities. Project Linchpin followed suit in November 2023. Private enterprises in industries as disparate as technology and food and beverage are beginning to charge their security professionals with solving for AI supply chain, as we discussed with Omar Santos of Cisco and Allan Friedman of CISA at a recent webinar.
Our work, in partnership with The Fletcher School at Tufts University, will focus on the feasibility of automatically generating AIBOMs at scale for Air Force components. We applaud the Air Force’s recognition that AIBOMs are one of the most pressing and urgent challenges within the DAF today, and we look forward to delivering AI supply chain security to the Air Force, elsewhere in the US Government, and throughout the world.
To read more about AIBOMs, access our whitepaper here, visit our GitHub, or contact us at info@manifestcyber.com.
