Bach, T. et al. 2026. Rethinking Deep Alignment Through the Lens of Incomplete Safety Learning. Proceedings of the AAAI Conference on Artificial Intelligence. 40, 36 (Mar. 2026), 30005–30012. DOI:https://doi.org/10.1609/aaai.v40i36.40248.