This project proposes to investigate the vulnerability of Analog Artificial Neural Network (AANN) integrated circuits (ICs) to radiation-based attacks aiming to distort the model implemented by these networks. We plan to leverage an existing AANN experimentation platform [1], which the PI designed and implemented in the past, to characterize the impact of radiation on various parts of the AANN IC. Building on the PI’s expertise and infrastructure for testing analog/mixed-signal ICs, as well as the Co-PI’s expertise in radiation-tolerance and infrastructure for conducting harsh environment and radiation effect experiments, we will develop and assess the effectiveness of an array of radiation-based attack scenarios on our AANN IC, seeking to understand its susceptibility and to inform future design modifications for radiation hardening of AANN ICs.