Skip to content
Learn Motion
ExploreHow it worksMembership
Log in
Learn Motion

Can AI Discover New Drugs? A High-Level Overview

1Why Drug Discovery Is Hard, and Where AI Fits2How AI Learns From Molecules and Proteins3Finding and Validating a Biological Target4Designing Molecules: Generative AI and Virtual Screening5From Hit to Lead: Optimizing Properties With AI6What AI Still Cannot Do7Judging the Claims: Real Successes, Failures, and Open Questions
Judging the Claims: Real Successes, Failures, and Open Questions

What Remains Genuinely Unsettled

4 / 4
Notice the wording of these four questions. None of them asks whether AI is useful — that is settled in specific, bounded ways. Each one asks whether the field can demonstrate that usefulness in terms that survive scrutiny. The first question is the one to watch over the coming years: do AI-designed molecules actually succeed more often in human testing, or do they mainly get us to the same outcome faster? We cannot answer it yet because the early programs are only now reaching that stage, and the numbers are small. The other three are open for a different reason — the evidence needed to close them does not exist yet. That distinction matters. An open question is not a refutation, and it is not a promise. It is a place where the honest answer today is that we do not know.
0:00 / 0:00

Four questions still without settled answers

  • Do AI-designed molecules succeed more often in human testing than conventionally designed ones, or do they mainly reach the same outcome faster and more cheaply?
  • How can we check whether a model has learned a causal mechanism rather than a correlation that happens to hold in the training data?
  • How should the field generate the missing data in underrepresented tissues and patient groups, given that better models cannot create measurements that were never taken?
  • What evidence from a model will regulators and clinicians accept when the prediction cannot be inspected or explained?

Why these are open rather than merely unanswered

Each question is open because the evidence needed to close it does not yet exist, not because no one has looked. The first requires years of clinical programs to mature. The second requires methods that are still being developed. The third requires data collection that is slow and costly by nature. The fourth depends on regulatory practice that is evolving alongside the technology. Recognizing this prevents the mistake of treating a current limitation as a permanent one, or a current claim as a settled result.

Previous4 / 4Complete

Learn Motion

Generate a course. Learn it properly.

Operated by Wuhan Daoyin Technology Co., Ltd.

Contact: [email protected]
Privacy PolicyTerms of Service

© 2026 Learn Motion