The purpose of this phase i project is to address critical gaps in AI safety and explainability

Objectsecurity LLC received a $106,412.26 cooperative agreement from the National Institute of Standards and Technology (NIST) under the Science, Technology, Business and/or Education Outreach program (CFDA 11.620), effective August 1, 2025 through January 31, 2026. The Phase I project addresses critical gaps in Artificial Intelligence (AI) safety and explainability by developing a novel technology that enhances AI content moderation using linear layer conductance techniques. The solution is designed to increase analysis efficiency without significant computational overhead while providing deeper insights into token relationships and patterns that influence AI outputs, enabling developers and end users to precisely moderate AI-generated content and prevent harmful outputs.

The recipient will execute activities including analysis and requirements research, proof-of-concept development, and production of reports to validate the proposed approach. No subrecipient activities are planned under this award. The technology is intended to ensure safer and more trustworthy AI systems, with primary beneficiaries in the military and healthcare sectors where AI exploitation poses significant operational and safety risks. Work will be performed at the recipient's facility in San Diego, California.

Activities

Analysis and requirementsresearch develop proof of conceptproduce reportsexpected outcomes: by automating complex analysis and enhancing model explainability, objectsecurity anticipates that its proposed technology will ensure safer and more trustworthy AI systems. Intended beneficiaries include: military and healthcare sectors.

Subrecipient activities

The recipient does not intend to subaward funds.