OpenTox 2.0 - An Evidence-Centered Reference Architecture for Next Generation Predictive Toxicology and AI-Assisted Decision Support

Predictive toxicology has evolved dramatically since the original OpenTox framework was introduced in 2010. New Approach Methodologies (NAMs), mechanistic models, FAIR knowledge infrastructures and artificial intelligence (AI) offer unprecedented opportunities to advance human and environmental safety assessment. Yet a fundamental challenge remains: how can we transform increasingly diverse scientific information and computational predictions into trustworthy evidence supporting transparent and accountable decisions? This lecture introduces OpenTox 2.0, a proposed evidence-centered reference architecture that extends the original OpenTox principles of interoperability, openness and reproducibility towards contextual scientific evidence assessment and AI-assisted decision support. We will examine its six conceptual layers, canonical information model and five proposed composable conformance profiles, showing how they connect scientific resources, computational services, workflows, evidence assessments and accountable decisions. A practical illustration will explore conflicting in silico predictions and demonstrate how OpenModel documentation, evidence-based ASPA workflows and OpenTox 2.0 architectural concepts can support model evaluation, uncertainty assessment and transparent weight-of-evidence reasoning. Applications to Next Generation Risk Assessment (NGRA) and Safe and Sustainable by Design (SSbD) will also be considered. Looking ahead, we will explore the opportunities and responsibilities associated with increasingly capable AI systems. Drawing on Isaac Asimov's original Three Laws of Robotics and a recently proposed extension comprising a foundational Zeroth Law and eleven principles, we will consider how truthfulness, epistemic humility, human oversight and responsible governance can inform trustworthy scientific AI. Finally, we will announce the establishment of the OpenTox Evidence Science Working Group, inviting the international community to collaborate on elaborating, implementing and independently testing OpenTox 2.0. Our ambition is to move beyond interoperable models towards evidence-centered scientific systems that enable more transparent, reproducible and defensible decisions for a safer and more sustainable future. References 1) Hardy, B., & Ali, A. (2026). OpenTox 2.0: A Perspective on the Principles for Predictive Toxicology and Risk Assessment Applications in the Era of Integrated New Approach Methods, Computational Modelling and Artificial Intelligence. Zenodo. https://doi.org/10.5281/zenodo.19458289 2) Hardy, B. (2026). Beyond Asimov's Three Laws: Eleven Principles for AI, Humanity and a Flourishing Future. Zenodo. https://doi.org/10.5281/zenodo.24218481 3) Barry Hardy et al. (2010). Collaborative development of predictive toxicology applications. J Cheminform 2, 7. https://doi.org/10.1186/1758-2946-2-7 4) Barry Hardy et al. (2012a). A toxicology ontology roadmap. ALTEX 29, 129–137. https://doi.org/10.14573/altex.2012.2.129

Authors

Institutions

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-16
DOI
https://doi.org/10.5281/zenodo.24285475
Primary Topic
Scientific Computing and Data Management
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

OpenTox 2.0 - An Evidence-Centered Reference Architecture for Next Generation Predictive Toxicology and AI-Assisted Decision Support

Barry J. Hardy
Zenodo (CERN European Organization for Nuclear Research)
Scientific Computing and Data Management
article

OpenTox 2.0 - An Evidence-Centered Reference Architecture for Next Generation Predictive Toxicology and AI-Assisted Decision Support

Barry J. Hardy
article en

Abstract

Predictive toxicology has evolved dramatically since the original OpenTox framework was introduced in 2010. New Approach Methodologies (NAMs), mechanistic models, FAIR knowledge infrastructures and artificial intelligence (AI) offer unprecedented opportunities to advance human and environmental safety assessment. Yet a fundamental challenge remains: how can we transform increasingly diverse scientific information and computational predictions into trustworthy evidence supporting transparent and accountable decisions? This lecture introduces OpenTox 2.0, a proposed evidence-centered reference architecture that extends the original OpenTox principles of interoperability, openness and reproducibility towards contextual scientific evidence assessment and AI-assisted decision support. We will examine its six conceptual layers, canonical information model and five proposed composable conformance profiles, showing how they connect scientific resources, computational services, workflows, evidence assessments and accountable decisions. A practical illustration will explore conflicting in silico predictions and demonstrate how OpenModel documentation, evidence-based ASPA workflows and OpenTox 2.0 architectural concepts can support model evaluation, uncertainty assessment and transparent weight-of-evidence reasoning. Applications to Next Generation Risk Assessment (NGRA) and Safe and Sustainable by Design (SSbD) will also be considered. Looking ahead, we will explore the opportunities and responsibilities associated with increasingly capable AI systems. Drawing on Isaac Asimov's original Three Laws of Robotics and a recently proposed extension comprising a foundational Zeroth Law and eleven principles, we will consider how truthfulness, epistemic humility, human oversight and responsible governance can inform trustworthy scientific AI. Finally, we will announce the establishment of the OpenTox Evidence Science Working Group, inviting the international community to collaborate on elaborating, implementing and independently testing OpenTox 2.0. Our ambition is to move beyond interoperable models towards evidence-centered scientific systems that enable more transparent, reproducible and defensible decisions for a safer and more sustainable future. References 1) Hardy, B., & Ali, A. (2026). OpenTox 2.0: A Perspective on the Principles for Predictive Toxicology and Risk Assessment Applications in the Era of Integrated New Approach Methods, Computational Modelling and Artificial Intelligence. Zenodo. https://doi.org/10.5281/zenodo.19458289 2) Hardy, B. (2026). Beyond Asimov's Three Laws: Eleven Principles for AI, Humanity and a Flourishing Future. Zenodo. https://doi.org/10.5281/zenodo.24218481 3) Barry Hardy et al. (2010). Collaborative development of predictive toxicology applications. J Cheminform 2, 7. https://doi.org/10.1186/1758-2946-2-7 4) Barry Hardy et al. (2012a). A toxicology ontology roadmap. ALTEX 29, 129–137. https://doi.org/10.14573/altex.2012.2.129

Zenodo (CERN European Organization for Nuclear Research)
Edelweiss Connect (Switzerland) (CH)
Scientific Computing and Data Management
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon