RxSafeBench: A reality check on AI medication safety

Kari Jaaskelainen

07 Nov 2025 — 1 min read

AI assistants are getting better at healthcare tasks—but are they safe when it comes to medications? A new study introduces RxSafeBench, a realistic test to see whether large language models (LLMs) can spot medication risks during simulated doctor-patient chats.

Using a curated RxRisk DB (6,725 contraindications, 28,781 drug interactions, 14,906 indication-drug pairs) and a two-stage clinical review, the team built 2,443 high-quality consultation scenarios with embedded risks. Leading open and proprietary LLMs were asked to choose safe treatments based on each patient's context.

Result: Current LLMs often miss contraindications and interactions—especially when the danger is implied, not stated.
Implication: Better prompting and task-specific tuning help, but safety gaps remain.
Contribution: RxSafeBench offers the first comprehensive benchmark for medication safety in LLMs.

Why it matters: Reliable AI should protect patients from harmful prescriptions, not just provide quick answers. This benchmark gives researchers and developers a common yardstick to build—and verify—safer clinical AI.

Paper: http://arxiv.org/abs/2511.04328v1

Paper: http://arxiv.org/abs/2511.04328v1

Register: https://www.AiFeta.com

#AI #Healthcare #PatientSafety #MedicationSafety #LLMs #NLP #ClinicalAI #Benchmark #Pharmacovigilance

Automating GDPR Compliance: A Roadmap for Companies and Law Firms

GDPR compliance is more than checkboxes. A new roadmap from the Privatech project shows how automation and machine learning can help companies and law firms assess—and even generate—privacy compliance. * Shift the focus to data processors’ real workflows: drafting policies, mapping data uses, documenting decisions. * Break compliance into machine-ready

FPGAs for Faster, Leaner Deep Learning: A Review of CNN Accelerators

Deep learning drives image search, robots, and medical scans. Most systems lean on CPUs and GPUs. This review asks: what if we run convolutional neural networks (CNNs) on FPGAs—reconfigurable chips you can tailor to the model? * Why FPGAs: custom dataflows, low latency, and strong energy efficiency—great for cameras,

Dynamic-K: Recommendations That Know When to Stop

Most apps show a fixed number of “top” items—say 10 movies or 20 products—assuming there are always enough good options. But that’s not always true: sometimes there are few relevant items, or some users are extra picky. The result? Filler recommendations. Dynamic-K flips the script. Instead of

Teaching chatbots to stop contradicting themselves (DECODE)

Teaching chatbots to stop contradicting themselves Ever had a bot say one thing, then the opposite a few turns later? This study introduces DECODE—a new task and dataset for spotting contradictions in everyday conversations, drawn from both human-human and human-bot chats. * New data beats existing natural language inference (NLI)

Read more

Automating GDPR Compliance: A Roadmap for Companies and Law Firms

FPGAs for Faster, Leaner Deep Learning: A Review of CNN Accelerators

Dynamic-K: Recommendations That Know When to Stop

Teaching chatbots to stop contradicting themselves (DECODE)