AI in Drug Discovery

A cross-cutting theme across SDDS 2026: where AI is already load-bearing, where it’s hype, and what constrains its impact.

Session I Perspectives

Where AI works today (table stakes)

• Regulatory submissions, trial-site optimization, enrollment, supply chain
• AI-enabled DMTA (design-make-test-analyze) cycles
• Real-time portfolio management (Sanofi’s agentic AI workflow)
• Single-cell immunology analysis (tractable questions that were impossible 5 years ago)

Where AI hasn’t delivered yet

• End-to-end connectivity from target ID through clinical pharmacology
• The AI-native biotech thesis: three-plus years on, timelines haven’t moved — 10 years from start to approval has held steady for decades

The binding constraint: data, not algorithms

• Akinsanya: “We can’t leverage AI/ML when the data isn’t in the dataset”
• Public chemistry and structure data are saturated
• An ML model learned which hospital a slide came from, not the biology — data provenance is the real issue
• The training data for dynamic biology mostly doesn’t exist yet

Risks

• Quigley’s worry: AI tools eroding critical thinking in early scientists who outsource the “what’s the next experiment?” question to a model

Cross-References

• Productivity paradox — AI is necessary but not sufficient
• Dynamic proteome — the data gap that limits AI’s reach
• Lab in the loop — Genentech’s iterative AI-experiment architecture