AI-Assisted Regulatory-Ready Workflows for Safe and Sustainable by Design (SSbD): From Fragmented Evidence to Mechanistically Interpretable Decision Support

Abstract Regulatory authorities such as the European Chemicals Agency are transitioning from reactive hazard assessment toward predictive, mechanistic and lifecycle-integrated regulatory science. This shift is driven by increasing material complexity (including advanced materials and nanoforms), the reduction of animal testing, and growing policy emphasis on Safe and Sustainable by Design (SSbD). However, the limiting factor for SSbD implementation is no longer scientific capability, but the lack of regulatory-ready evidence structures: integrated workflows, reproducibility frameworks, traceable metadata, and clearly defined contexts of use. This presentation introduces an integrated, AI-assisted, FAIR knowledge infrastructure supporting tiered SSbD workflows, developed within SSbD4CheM and operationalised through the SaferWorldbyDesign platform. Building on the Alternative Safety Profiling Algorithm (ASPA) framework (Leist et al., 2026), the approach combines NAM-based Integrated Approaches to Testing and Assessment (IATA), exposure and toxicokinetic modelling, life cycle sustainability assessment, and structured knowledge graphs into fully traceable regulatory evidence packages. AI is deployed not as a black-box decision maker, but as a transparent evidence integrator supporting mechanistic interpretation, uncertainty analysis, and applicability domain definition. The framework directly addresses key regulatory barriers: (i) fragmentation of experimental and modelling evidence, (ii) lack of standardised regulatory-ready data structures, (iii) insufficient reproducibility and validation documentation, and (iv) absence of integrated interpretation across hazard, exposure and sustainability dimensions. Demonstrated use cases illustrate how structured metadata, model provenance, and documented decision gates enable earlier safety-informed material design and regulatory confidence. By embedding mechanistically interpretable NAM data, computational modelling, and lifecycle sustainability assessment within reproducible, FAIR workflows, this approach supports regulators in making high-confidence decisions under uncertainty while accelerating innovation. The strategic vision is a transition toward predictive regulatory science in which SSbD evidence is generated continuously across the innovation lifecycle, from early design optimisation through regulatory submission and post-market learning. Context Invited presentation, 11 March 2026, Anthos 26 conference session on Solutions Session 3: Project solutions to the needs of Regulators (ECHA) https://www.bnn.at/anthos26_session3-regulators/ Reference Leist, M., Tangianu, S., Affourtit, F., Braakhuis, H., Colbourne, J., Cöllen, E., Dreser, N., Escher, S.E., Gardner, I., Hahn, S., Hardy, B., Herzler, M., Islam, B., Kamp, H., Magel, V., Marx-Stoelting, P., Moné, M.J., Lundquist, P., Ottenbros, I., Ouedraogo, G., Pallocca, G., van de Water, B., Vinken, M., White, A., Pastor, M. and Luijten, M. (2026) ‘An Alternative Safety Profiling Algorithm (ASPA) to transform next generation risk assessment into a structured and transparent process’, ALTEX, 43(1), pp. 158–175.