Enter Note Done
Go to previous page in this tab
Event
  • Featured Session
  • Bookmark Icon
  • SOT/EUROTOX Debate: Artificial Intelligence in Toxicology Represents a Transformative Revolution Rather than a Cause for Caution
  • Schedule
    Notes
  • Stars at Night Ballroom B1‚ Convention Center
    Set Timezone


Chair: Barbara L.F. Kaplan, Mississippi State University
Co-Chair: Hilmi Orhan, Ege Üniversitesi, Turkey


The annual SOT/EUROTOX debate features leading toxicologists presenting contrasting perspectives on a controversial or pressing issue in toxicology. The debate began in the early 1990s and has been a highly anticipated event ever since. This year, the debaters will address the proposition “Artificial Intelligence in Toxicology Represents a Transformative Revolution Rather than a Cause for Caution.”

The debaters will introduce ways that AI can be used in toxicology, discuss AI as new approaches and models, and how these models improve on current assays. Critical to this debate are considerations of ethical issues surrounding use, development of new approaches and models, accessibility, effective communication of use and applicability, and translatability, as well as use of natural resources and energy consumption required to maintain its use. Relevant questions might include: (1) Is AI currently reliable enough to replace or substantially reduce conventional toxicology assays? (2) Does AI improve toxicological prediction and decision-making more than it introduces new sources of error, bias, and uncertainty? (3) Should AI-generated evidence be accepted for regulatory toxicology decisions without confirmatory experimental testing? (4) Can AI models provide mechanistic insight and biological interpretability, or do they primarily generate predictions without sufficient explanation? (5) When an AI model produces an inaccurate, irreproducible, or obsolete result, who is accountable, and how should that evidence be corrected or withdrawn? (6) Will AI democratize toxicology by expanding access to advanced prediction tools, or will it deepen inequities between well-resourced institutions and others? (7) Is AI a true new approach methodology, or is it primarily a supporting tool that still depends on conventional toxicology data and experimental validation? (8) Do the potential benefits of AI, such as faster hazard identification, improved chemical prioritization, and reduced animal use, outweigh its energy, water, hardware, and data-storage costs? (9) How can AI users better disseminate their use of AI in publishing, data analysis and interpretation, and model development? (10) How can we improve on AI tool accessibility to toxicologists and others?

In addition to inclusion as a Featured Session at this meeting, this debate will take place again (with the debaters taking the reverse positions) in Gothenburg, Sweden, during EUROTOX 2027, September 5–8, 2027.

Session and Activity Type
Presentations