Autonomous AI Agents Can Outperform Physicians. That’s Not the Hard Part.
MIRA beat physicians on diagnostic accuracy. AMIE matched them across 100 cases. Now comes the real test: whether regulated...
AI in Clinical Trials
Nature Medicine’s June 2026 Benchmark Study Reveals General-Purpose LLMs Outperform FDA-Cleared Clinical AI — and Exposes a Validation Gap Regulators Have Not Closed
A Nature Medicine study finds GPT-5.2, Gemini 3.1 Pro, and Claude Opus 4.6 outperform FDA-cleared clinical AI tools on...
Lilly’s Bet on Abridge Exposes the Recruitment Handoff Nobody Wants to Own
Eli Lilly invested in Abridge's AI recruitment tool — but the real question is whether sites can operationalize patient...
AI Site Selection Sounds Fast — Until You See What Happens After the Algorithm Picks Your Site
ASCO and Ryght AI are using algorithms to select CDK4/6 inhibitor trial sites. Here's what ops teams need to...
When AI Picks Your Site, Execution Still Happens in the Building
ASCO and Ryght AI are using machine learning to select breast cancer trial sites. Here's what that means for...
Veeva Unveils Falcon AI Platform and Agentic Authoring at 2026 Summit
At the 2026 Veeva R&D and Quality Summit in Copenhagen, Veeva Systems announced three major AI capabilities: Vault AI,...
When Bad Protocol Design Hits Your Site, AI Won’t Save You
Phesi finds fewer than 1 in 3 trial protocols connect to real patient data — meaning flawed assumptions reach...
When AI Scales Your Protocol, It Scales Every Flaw Inside It
Phesi's 2026 analysis found fewer than 1 in 3 trial protocols are linked to patient data—meaning AI tools are...
SPARK Just Ran 5,400-Patient Oncology Research Autonomously. Your Trial’s Biomarker Strategy Has Not Caught Up.
SPARK agentic AI validated prognostic biomarkers across 5,400 patients and 18 cancer types—here's what that means for trial enrichment...
FDA Real-Time Clinical Trials: Why AI Systems That Update Themselves Are Breaking Clinical Trial Design
FDA's April 2026 real-time clinical trial push exposes a fatal flaw: when AI updates itself mid-study, your protocol is...