AI in drug discovery — what it is, where we stand and the path forward

A Perspective in Nature Reviews Drug Discovery finds that despite a decade of intense interest, AI in drug discovery has shown disappointingly limited clinically relevant impact, with no evidence yet that AI-discovered drugs outperform traditional ones in trials. The authors, citing 2024 FDA approvals and clinical development success rates, attribute the gap to insufficient focus on clinical translation, underspecified models, and a 'technology push' rather than 'science pull' dynamic, and recommend shifting benchmarking from model validation to decision-making improvement.

Abstract Artificial intelligence AI in drug discovery has attracted increasing interest over the past decade. It is now time for a critical review of progress in the field: where did we advance — and where are we yet to see impact — when it comes to what matters in drug discovery, which is to deliver safer and more efficacious medicines to patients faster? Although a wide variety of AI methods have been developed, applied and benchmarked, evidence of their clinically relevant impact is, so far, disappointingly limited. In this Perspective we discuss potential reasons, including an insufficient focus on clinical translation during model development, difficulties with applying AI algorithms on conditional life science data, and insufficient problem definitions and the resulting underspecification of computational models for real-world use cases. ‘Technology push’ compared with ‘science pull’ is also likely to be an underlying factor, as well as the substantial time required to operationalize technical capabilities into systems that are sufficiently scaled and accessible for users. We provide recommendations for the development of AI in drug discovery with the aim of increasing its translational relevance. For example, benchmarking studies of AI tools in drug discovery need to move on from model validation and instead focus on their ability to improve decision making. This is a preview of subscription content, access via your institution https://wayf.springernature.com?redirect uri=https%3A%2F%2Fwww.nature.com%2Farticles%2Fs41573-026-01496-2 Access options Access Nature and 54 other Nature Portfolio journals Get Nature+, our best-value online-access subscription 27,99 € / 30 days cancel any time Subscribe to this journal Receive 12 print issues and online access 269,00 € per year only 22,42 € per issue Buy this article - Purchase on SpringerLink - Instant access to the full article PDF. 39,95 € Prices may be subject to local taxes which are calculated during checkout Similar content being viewed by others References Cireşan, D. C., Meier, U., Gambardella, L. M. & Schmidhuber, J. Deep, big, simple neural nets for handwritten digit recognition. Neural Comput. 22 , 3207–3220 2010 .Krizhevsky, A., Sutskever, I. & Hinton, G. E. ImageNet classification with deep convolutional neural networks. In Advances in Neural Information Processing Systems 25 NIPS 2012 1097–1105 Curran Associates, Inc., 2012 .Hochreiter, S. & Schmidhuber, J. Long short-term memory. Neural Comput. 9 , 1735–1780 1997 .Walters, W. P. & Barzilay, R. Critical assessment of AI in drug discovery. Expert. Opin. Drug Discov. 16 , 937–947 2021 .Rodriguez, A. et al. Unlocking the Potential of AI in Drug Discovery Boston Consulting Group Report, 2023 .Jayatunga, M. K. P., Xie, W., Ruder, L., Schulze, U. & Meier, C. AI in small-molecule drug discovery: a coming wave? Nat. Rev. Drug Discov. 21 , 175–176 2022 .Mullard, A. 2024 FDA approvals. Nat. Rev. Drug Discov. 24 , 75–82 2025 .Phares, S., Phillip, K. & Trusheim, M. Clinical development success rates for durable cell and gene therapies. Nat. Rev. Drug Discov. 24 , 329–330 2025 .Wouters, O. J. et al. Differential legal protections for biologics vs small-molecule drugs in the US. J. Am. Med. Assoc. 332 , 2101–2108 2024 .Jayatunga, M., Ayers, M., Bruens, L., Jayanth, D. & Meier, C. How successful are AI-discovered drugs in clinical trials? A first analysis and emerging lessons. Drug Discov. Today 29 , 104009 2024 .Lowe, D. AI drugs so far. In the Pipeline https://www.science.org/content/blog-post/ai-drugs-so-far https://www.science.org/content/blog-post/ai-drugs-so-far 2024 .Pitt, W. R. et al. Real-world applications and experiences of AI/ML deployment for drug discovery. J. Med. Chem. 68 , 851–859 2025 .Marshall, S. et al. Model-Informed drug discovery and development: current industry good practice and regulatory expectations and future perspectives. CPT Pharmacometrics Syst. Pharmacol. 8 , 87–96 2019 .Madhavan, S. & Shaywitz, D. A. AI: an essential tool for managing the burgeoning complexity of clinical development in pharmaceutical R&D. Drug Discov. Today 30 , 104271 2025 .Wodak, S. J., Vajda, S., Lensink, M. F., Kozakov, D. & Bates, P. A. Critical assessment of methods for predicting the 3D structure of proteins and protein complexes. Annu. Rev. Biophys. 52 , 183–206 2023 .Vázquez Torres, S. et al. De novo designed proteins neutralize lethal snake venom toxins. Nature 639 , 225–231 2025 .Du, H. et al. Targeting peptide antigens using a multiallelic MHC I-binding system. Nat. Biotechnol. 43 , 1683–1693 2024 .De Jonghe, J. et al. A community effort to track commercial single-cell and spatial ’omic technologies and business trends. Nat. Biotechnol. 42 , 1017–1023 2024 .Göller, A. H. et al. Bayer’s in silico ADMET platform: a journey of machine learning over the past two decades. Drug Discov. Today 25 , 1702–1709 2020 .Scannell, J. W. et al. Predictive validity in drug discovery: what it is, why it matters and how to improve it. Nat. Rev. Drug Discov. 21 , 915–931 2022 .Clark, D. E. What has computer-aided molecular design ever done for drug discovery? Expert Opin. Drug Discov. 1 , 103–110 2006 .Bender, A. & Cortés-Ciriano, I. Artificial intelligence in drug discovery: what is realistic, what are illusions? Part 1: ways to make an impact, and why we are not there yet. Drug Discov. Today 26 , 511–524 2021 .Paul, S. et al. How to improve R&D productivity: the pharmaceutical industry’s grand challenge. Nat. Rev. Drug Discov. 9 , 203–214 2010 .Dreiman, G. H. S. et al. Changing the HTS Paradigm: AI-driven iterative screening for hit finding. SLAS Discov. 26 , 257–262 2021 .Morgan, P. et al. Impact of a five-dimensional framework on R&D productivity at AstraZeneca. Nat. Rev. Drug Discov. 17 , 167–181 2018 .Fernando, K. et al. Achieving end-to-end success in the clinic: Pfizer’s learnings on R&D productivity. Drug Discov. Today 27 , 697–704 2022 .Wong, C. H., Siah, K. W. & Lo, A. W. Estimation of clinical trial success rates and related parameters. Biostatistics 20 , 273–286 2019 .Minikel, E. V. et al. Refining the impact of genetic evidence on clinical success. Nature 629 , 624–629 2024 .Bender, A. & Cortes-Ciriano, I. Artificial intelligence in drug discovery: what is realistic, what are illusions? Part 2: a discussion of chemical and biological data. Drug Discov. Today 26 , 1040–1052 2021 .Ortiz de Montellano, P. R. Cytochrome P450-activated prodrugs. Future Med. Chem. 5 , 213–228 2013 .Rodriguez-Antona, C. & Ingelman-Sundberg, M. Cytochrome P450 pharmacogenetics and cancer. Oncogene 25 , 1679–1691 2006 .Zevin, S. & Benowitz, N. L. Drug interactions with tobacco smoking. An update. Clin. Pharmacokinet. 36 , 425–438 1999 .Zimmermann, M. et al. Separating host and microbiome contributions to drug pharmacokinetics and toxicity. Science 363 , eaat9931 2019 .Safikhani, Z. et al. Revisiting inconsistency in large pharmacogenomic studies. F1000Research 5 , 2333 2016 .Samad, S. S., Schwartz, J. M. & Francavilla, C. Functional selectivity of receptor tyrosine kinases regulates distinct cellular outputs. Front. Cell Dev. Biol. 11 , 1348056 2024 .Zhang, Z., Pan, Q., Lu, M. & Zhao, B. Intermediate endpoints as surrogates for outcomes in cancer immunotherapy: a systematic review and meta-analysis of phase 3 trials. eClinicalMedicine 63 , 102156 2023 .Nogales, C. et al. Network medicine-based unbiased disease modules for drug and diagnostic target identification in ROSopathies. Handb. Exp. Pharmacol. 264 , 49–68 2021 .Grotzinger, A. D. et al. Mapping the genetic landscape across 14 psychiatric disorders. Nature https://doi.org/10.1038/s41586-025-09820-3 https://doi.org/10.1038/s41586-025-09820-3 2025 .Polishchuk, P. G., Madzhidov, T. I. & Varnek, A. Estimation of the size of drug-like chemical space based on GDB-17 data. J. Comput. Aided Mol. Des. 27 , 675–679 2013 .Balestriero, R. et al. Learning in high dimension always amounts to extrapolation. Preprint at https://arxiv.org/abs/2110.09485 https://arxiv.org/abs/2110.09485 2021 .Sheridan, R. P. Time-split cross-validation as a method for estimating the goodness of prospective prediction. J. Chem. Inf. Model. 53 , 783–790 2013 .Lee, K., Moldagulov, G. & Grzybowski, B. A. The fragility of bioactivity prediction: rigorous dataset splits expose the illusion of ML accuracy. Chem. Eur. J. 15 , e71208 2026 .Kapoor, S. & Narayanan, A. Leakage and the reproducibility crisis in machine-learning-based science. Patterns 4 , 100804 2023 .Rodgers, S., Glen, R. C. & Bender, A. Characterizing bitterness: identification of key structural features and development of a classification model. J. Chem. Inf. Model. 46 , 569–576 2006 .Fierro, F., Giorgetti, A., Carloni, P., Meyerhof, W. & Alfonso-Prieto, M. Dual binding mode of “bitter sugars” to their human bitter taste receptor target. Sci. Rep. 9 , 8437 2019 .Huang, G., Lv, M., Hu, J., Huang, K. & Xu, H. Glycosylation and activities of natural products. Mini Rev. Med. Chem. 16 , 1013–1016 2016 .D’Amour, A. et al. Underspecification presents challenges for credibility in modern machine learning. J. Mach. Learn. Res. 23 , 1–61 2022 .Bender, A. et al. Evaluation guidelines for machine learning tools in the chemical sciences. Nat. Rev. Chem. 6 , 428–442 2022 .Church, K. & Kordoni, V. Emerging trends: SOTA-chasing. Nat. Lang. Eng. 28 , 249–269 2020 .Peters, B., Brenner, S. E., Wang, E., Slonim, D. & Kann, M. G. Putting benchmarks in their rightful place: the heart of computational biology. PLoS Comput. Biol. 14 , e1006494 2018 .Moult, J., Pedersen, J. T., Judson, R. & Fidelis, K. A large-scale experiment to assess protein structure prediction methods. Proteins 23 , 2–5 1995 .Newman, J. et al. Practical aspects of the SAMPL challenge: providing an extensive experimental data set for the modeling community. J. Biomol. Screen. 14 , 1245–1250 2009 .Ackloo, S. et al. CACHE Critical Assessment of Computational Hit-finding Experiments : a public–private partnership benchmarking initiative to enable the development of computational methods for hit-finding. Nat. Rev. Chem. 6 , 287–295 2022 .Lander, E. S. et al. Initial sequencing and analysis of the human genome. Nature 409 , 860–921 2001 .Venter, J. C. et al. The sequence of the human genome. Science 291 , 1304–1351 2001 .Reiss, T. Drug discovery of the future: the implications of the human genome project. Trends Biotechnol. 19 , 496–499 2001 .Scannell, J. W. & Bosley, J. When quality beats quantity: decision theory, drug discovery, and the reproducibility crisis. PLoS ONE 11 , e0147215 2016 .Heyndrickx, W. et al. MELLODDY: cross-pharma federated learning at unprecedented scale unlocks benefits in QSAR without compromising proprietary information. J. Chem. Inf. Model. 64 , 2331–2344 2024 .Jumper, J. et al. Highly accurate protein structure prediction with AlphaFold. Nature 596 , 583–589 2021 .Liebeschuetz, J., Hennemann, J., Olsson, T. & Groom, C. R. The good, the bad and the twisted: a survey of ligand geometry in protein crystal structures. J. Comput. Aided Mol. Des. 26 , 169–183 2012 .Markowetz, F. All models are wrong and yours are useless: making clinical prediction models impactful for patients. npj Precis. Oncol. 8 , 54 2024 .Arnold, C. Inside the nascent industry of AI-designed drugs. Nat. Med. 29 , 1292–1295 2023 .Ackoff, R. L. The future of operational research is past. J. Oper. Res. Soc. 30 , 93–104 1979 .Kearnes, S. Pursuing a prospective perspective. Trends Chem. 3 , 77–79 2021 .Wellnitz, J. et al. One size does not fit all: revising traditional paradigms for assessing accuracy of QSAR models used for virtual screening. J. Cheminform. 17 , 7 2025 .Seal, S. et al. Machine learning for toxicity prediction using chemical structures: pillars for success in the real world. Chem. Res. Toxicol. 38 , 759–807 2025 .Obrezanova, O. et al. Prediction of in vivo pharmacokinetic parameters and time-exposure curves in rats using machine learning from the chemical structure. Mol. Pharm. 19 , 1488–1504 2022 .Miljković, F. et al. Machine learning models for human in vivo pharmacokinetic parameters with in-house validation. Mol. Pharm. 18 , 4520–4530 2021 .Walters, P. We need better benchmarks for machine learning in drug discovery. Practical Cheminformatics http://practicalcheminformatics.blogspot.com/2023/08/we-need-better-benchmarks-for-machine.html http://practicalcheminformatics.blogspot.com/2023/08/we-need-better-benchmarks-for-machine.html 2023 .Steiner, J. SOTA seeking — a knife fight in a phone booth. Techbio< B iotech https://biotechbio.substack.com/p/sota-seeking-a-knife-fight-in-a-phone https://biotechbio.substack.com/p/sota-seeking-a-knife-fight-in-a-phone 2023 .Ott, S. et al. Mapping global dynamics of benchmark creation and saturation in artificial intelligence. Nat. Commun. 13 , 6793 2022 .Youssef, A. et al. External validation of AI models in health should be replaced with recurring local validation. Nat. Med. 29 , 2686–2687 2023 .Del Rosario, Z., Rupp, M., Kim, Y., Antono, E. & Ling, J. Assessing the frontier: active learning, model accuracy, and multi-objective candidate discovery and optimization. J. Chem. Phys. 153 , 024112 2020 .Kalliokoski, T., Kramer, C., Vulpetti, A. & Gedeck, P. Comparability of mixed IC 50data – a statistical analysis. PLoS ONE 8 , e61007 2013 .Landrum, G. A. & Riniker, S. Combining IC 50or K ivalues from different sources is a source of significant noise. J. Chem. Inf. Model. 64 , 1560–1567 2024 .Cortés-Ciriano, I. & Bender, A. How consistent are publicly reported cytotoxicity data? Large-scale statistical analysis of the concordance of public independent cytotoxicity measurements. ChemMedChem 11 , 57–71 2016 .Brown, S. P., Muchmore, S. W. & Hajduk, P. J. Healthy skepticism: assessing realistic model performance. Drug Discov. Today 14 , 420–427 2009 .Crusius, D., Cipcigan, F. & Biggin, P. C. Are we fitting data or noise? Analysing the predictive power of commonly used datasets in drug-, materials-, and molecular-discovery. Faraday Discuss. 256 , 304–321 2025 .van Tilborg, D., Alenicheva, A. & Grisoni, F. Exposing the limitations of molecular machine learning with activity cliffs. J. Chem. Inf. Model. 62 , 5938–5951 2022 .Ahdritz, G. et al. OpenFold: retraining AlphaFold2 yields new insights into its learning mechanisms and capacity for generalization. Nat. Methods 21 , 1514–1524 2024 .Kryshtafovych, A., Schwede, T., Topf, M., Fidelis, K. & Moult, J. Critical assessment of methods of protein structure prediction CASP —round XIV. Proteins 89 , 1607–1617 2021 .Scardino, V., Di Filippo, J. I. & Cavasotto, C. N. How good are AlphaFold models for docking-based virtual screening? iScience 26 , 105920 2023 .Lyu, J. et al. AlphaFold2 structures guide prospective ligand discovery. Science 384 , eadn6354 2024 .Karelina, M., Noh, J. J. & Dror, R. O. How accurately can one predict drug binding modes using AlphaFold models? eLife 12 , RP89386 2023 .Masters, M. R., Mahmoud, A. H. & Lill, M. A. Investigating whether deep learning models for co-folding learn the physics of protein–ligand interactions. Nat. Commun. 16 , 8854 2025 .Star-studded, A. I. Biotech launch. Nat. Biotechnol. 42 , 689 2024 .Du, Y. et al. Machine learning-aided generative molecular design. Nat. Mach. Intell. 6 , 589–604 2024 .Thomas, M. et al. Identification of nanomolar adenosine A2A receptor ligands using reinforcement learning and structure-based drug design. Nat. Commun. 16 , 5485 2025 .Atomwise AIMS Program AI is a viable alternative to high throughput screening: a 318-target study. Sci. Rep. 14 , 7526 2024 .Xu, Z. et al. A generative AI-discovered TNIK inhibitor for idiopathic pulmonary fibrosis: a randomized phase 2a trial. Nat. Med. 31 , 2602–2610 2025 .Ren, F. et al. A small-molecule TNIK inhibitor targets fibrosis in preclinical and clinical models. Nat. Biotechnol. 43 , 63–75 2025 .Lenselink, E. B. et al. Beyond the hype: deep neural networks outperform established methods using a ChEMBL bioactivity benchmark set. J. Cheminform. 9 , 45 2017 .Mayr, A. et al. Large-scale comparison of machine learning methods for drug target prediction on ChEMBL. Chem. Sci. 9 , 5441–5451 2018 .Janela, T. & Bajorath, J. Simple nearest-neighbour analysis meets the accuracy of compound potency predictions using complex machine learning models. Nat. Mach. Intell. 4 , 1246–1255 2022 .Robinson, M. C., Glen, R. C. & Lee, A. A. Validating the validation: reanalyzing a large-scale comparison of deep learning and machine learning models for bioactivity prediction. J. Comput. Aided Mol. Des. 34 , 717–730 2020 .Lane, T. R. et al. Bioactivity comparison across multiple machine learning algorithms using over 5000 datasets for drug discovery. Mol. Pharm. 18 , 403–415 2021 .Ferreira, F. J. & Carneiro, A. S. AI-driven drug discovery: a comprehensive review. ACS Omega 10 , 23889–23903 2025 .Eken, B. et al. A multivocal review of MLOps practices, challenges and open issues. ACM Comput. Surv. 58 , 1–35 2025 .Bray, M. A. et al. Cell Painting, a high-content image-based assay for morphological profiling using multiplexed fluorescent dyes. Nat. Protoc. 11 , 1757–1774 2016 .Seal, S. et al. Cell Painting: a decade of discovery and innovation in cellular imaging. Nat. Methods 22 , 254–268 2025 .Pruteanu, L. L. & Bender, A. Using transcriptomics and cell morphology data in drug discovery: the long road to practice. ACS Med. Chem. Lett. 14 , 386–395 2023 .Chandrasekaran, S. N. et al. Morphological map of under- and overexpression of genes in human cells. Nat. Methods 22 , 1742–1752 2025 .Celik, S. et al. Building, benchmarking, and exploring perturbative maps of transcriptional and morphological data. PLoS Comput. Biol. 20 , e1012463 2024 .Chen, D. et al. A combined AI and cell biology approach surfaces targets and mechanistically distinct Inflammasome inhibitors. iScience 27 , 111404 2024 .Trapotsi, M. A. et al. Cell morphological profiling enables high-throughput screening for PROteolysis TArgeting Chimera PROTAC phenotypic signature. ACS. Chem. Biol. 17 , 1733–1744 2022 .Keenan, A. B. et al. Connectivity mapping: methods and applications. Annu. Rev. Biomed. Data Sci. 2 , 69–92 2019 .Ewald, J. D. et al. Cell Painting for cytotoxicity and mode-of-action analysis in primary human hepatocytes. Cell Syst. 17 , 101566 2026 .Platani, M. et al. Screening for variable drug responses using human iPSC cohorts. PLoS ONE 20 , e0323953 2025 .Roohani, Y. H. et al. Virtual cell challenge: toward a turing test for the virtual cell. Cell 188 , 3370–3374 2025 .Fu, H., Hardy, J. & Duff, K. E. Selective vulnerability in neurodegenerative diseases. Nat. Neurosci. 21 , 1350–1358 2018 .Summers, R. A. et al. Novel human iPSC models of neuroinflammation in neurodegenerative disease and regenerative medicine. Trends Immunol. 45 , 799–813 2024 .Okano, H. & Morimoto, S. iPSC-based disease modeling and drug discovery in cardinal neurodegenerative disorders. Cell Stem Cell 29 , 189–208 2022 .Morimoto, S. et al. Phase 1/2a clinical trial in ALS with ropinirole, a drug candidate identified by iPSC drug discovery. Cell Stem Cell 30 , 766–780 2023 .Atmaramani, R. et al. Deep learning analysis on images of iPSC-derived Motor Neurons Carrying fALS-genetics Reveals Disease-Relevant Phenotypes. Preprint at bioRxiv https://doi.org/10.1101/2024.01.04.574270 https://doi.org/10.1101/2024.01.04.574270 2024 .Zushin, P. H., Mukherjee, S. & Wu, J. C. FDA Modernization Act 2.0: transitioning beyond animal models with human cells, organoids, and AI/ML-based approaches. J. Clin. Invest. 133 , e175824 2023 .Yu, S. et al. Integrating inflammatory biomarker analysis and artificial-intelligence-enabled image-based profiling to identify drug targets for intestinal fibrosis. Cell Chem. Biol. 30 , 1169–1182 2023 .Ingber, D. E. Human organs-on-chips for disease modelling, drug development and personalized medicine. Nat. Rev. Genet. 23 , 467–491 2022 .Chen, B., Du, C., Wang, M., Guo, J. & Liu, X. Organoids as preclinical models of human disease: progress and applications. Med. Rev. 4 , 129–153 2024 .Calandrini, C. & Drost, J. Normal and tumor-derived organoids as a drug screening platform for tumor-specific drug vulnerabilities. STAR Protoc. 3 , 101079 2022 .Scalia, G. et al. Deep-learning-based virtual screening of antibacterial compounds. Nat. Biotechnol. https://doi.org/10.1038/s41587-025-02814-6 https://doi.org/10.1038/s41587-025-02814-6 2025 .Chungyoun, M. & Gray, J. Fitness landscape for antibodies 2: benchmarking reveals that protein AI models cannot yet consistently predict developability properties. Preprint at bioRxiv https://doi.org/10.64898/2025.12.27.696706 https://doi.org/10.64898/2025.12.27.696706 2025 .Lim, J. et al. Advances in single-cell omics and multiomics for high-resolution molecular profiling. Exp. Mol. Med. 56 , 515–526 2024 .Ota, M. et al. Causal modelling of gene effects from regulators to programs to traits. Nature 650 , 399–408 2025 .Sivanandan, S. et al. A pooled Cell Painting CRISPR screening platform enables de novo inference of gene function by self-supervised deep learning. Nat. Commun. 17 , 77 2025 .Ghiandoni, G. M. et al. Augmenting DMTA using predictive AI modelling at AstraZeneca. Drug Discov. Today 29 , 103945 2024 .Medcalf, M. et al. Overcoming DMTA cycle challenges: a unified AI-driven system for efficient drug design. Preprint at https://doi.org/10.26434/chemrxiv-2024-0z7g6-v2 https://doi.org/10.26434/chemrxiv-2024-0z7g6-v2 2025 .Ma, S. et al. Evolving drug discovery using AI, automation, and ASMS through an integrated D-preMTA-MTA strategy for target-focused library exploration Preprint at. Research Square https://doi.org/10.21203/rs.3.rs-3753964/v1 https://doi.org/10.21203/rs.3.rs-3753964/v1 2023 .Tang, Q. et al. AI-driven robotics laboratory identifies pharmacological TNIK inhibition as a potent senomorphic agent. Aging Dis. 17 , 432–451 2025 .Alves, V. M. et al. Curated data in — trustworthy in silico models out: the impact of data quality on the reliability of artificial intelligence models as alternatives to animal testing. Altern. Lab. Anim. 49 , 73–82 2021 .Saha, U. S. et al. Step forward cross validation for bioactivity prediction: out of distribution validation in drug discovery. Preprint at bioRxiv https://doi.org/10.1101/2024.07.02.601740 https://doi.org/10.1101/2024.07.02.601740 2024 .Brumfield, M. A. et al. Delivering regulatory impact from consortium-based projects. Nat. Rev. Drug Discov. 24 , 889–890 2025 .Scannell, J. et al. Diagnosing the decline in pharmaceutical R&D efficiency. Nat. Rev. Drug Discov. 11 , 191–200 2012 .Truebel, H. & Seidler, M. Mitigating bias in pharmaceutical R&D decision-making. Nat. Rev. Drug Discov. 21 , 874–875 2022 .Shaywitz, D. You have chosen poorly: why drug developers make bad decisions. Will technology help? Timmermann Report https://timmermanreport.com/2022/10/you-have-chosen-poorly-why-drug-developers-make-bad-decisions/ https://timmermanreport.com/2022/10/you-have-chosen-poorly-why-drug-developers-make-bad-decisions/ 2022 .Otamendi, F. J. & Sutil Martín, D. L. The emotional effectiveness of advertisement. Front. Psychol. 11 , 2088 2020 .Bieske, L., Zinner, M., Dahlhausen, F. & Trübel, H. Trends, challenges, and success factors in pharmaceutical portfolio management: cognitive biases in decision-making and their mitigating measures. Drug Discov. Today 28 , 103734 2023 .Peplow, M. Robot chemist sparks row with claim it created new materials. Nature https://doi.org/10.1038/d41586-023-03956-w https://doi.org/10.1038/d41586-023-03956-w 2023 .Cokol, M., Iossifov, I., Rodriguez-Esteban, R. & Rzhetsky, A. How many scientific papers should be retracted? EMBO Rep. 8 , 422–423 2007 .Lazebnik, Y. Can a biologist fix a radio? Or, what I learned while studying apoptosis. Cancer Cell 2 , 179–182 2002 .Anderson, P. W.More is different. Science 177 , 393–396 1972 .Frangogiannis, N. G. Why animal model studies are lost in translation. J. Cardiovasc. Aging 2 , 22 2022 .Kitano, H. Systems biology: a brief overview. Science 295 , 1662–1664 2002 .Hasin, Y., Seldin, M. & Lusis, A. Multi-omics approaches to disease. Genome Biol. 18 , 83 2017 .Ravarani, C. N. J. et al. Retrospective evaluation of human genetic evidence for clinical trial success using Mendelian randomization and machine learning. Preprint at medRxiv https://doi.org/10.64898/2026.02.19.26346536 https://doi.org/10.64898/2026.02.19.26346536 2026 .Lagunin, A., Stepanchikova, A., Filimonov, D. & Poroikov, V. PASS: prediction of activity spectra for biologically active substances. Bioinformatics 16 , 747–748 2000 .Lounkine, E. et al. Large-scale prediction and testing of drug activity on side-effect targets. Nature 486 , 361–367 2012 .Bender, A. et al. Analysis of pharmacology data and the prediction of adverse drug reactions and off-target effects from chemical structure. ChemMedChem 2 , 861–873 2007 .Kirchmair, J. et al. Predicting drug metabolism: experiment and/or computation? Nat. Rev. Drug Discov. 14 , 387–404 2015 .Davies, M. et al. Improving the accuracy of predicted human pharmacokinetics: lessons learned from the astrazeneca drug pipeline over two decades. Trends Pharmacol. Sci. 41 , 390–408 2020 .Schneckener, S. et al. Prediction of oral bioavailability in rats: transferring insights from in vitro correlations to deep machine learning models using in silico model outputs and chemical structure parameters. J. Chem. Inf. Model. 59 , 4893–4905 2019 .Handa, K. et al. Prediction of compound plasma concentration-time profiles in mice using random forest. Mol. Pharm. 20 , 3060–3072 2023 .Andrews-Morger, A., Reutlinger, M., Parrott, N. & Olivares-Morales, A. A machine learning framework to improve rat clearance predictions and inform physiologically based pharmacokinetic modeling. Mol. Pharm. 20 , 5052–5065 2023 .Obach, R. S., Lombardo, F. & Waters, N. J. Trend analysis of a database of intravenous pharmacokinetic parameters in humans for 670 drug compounds. Drug Metab. Dispos. 36 , 1385–1405 2008 .Bauer, J. et al. Ritonavir: an extraordinary example of conformational polymorphism. Pharm. Res. 18 , 859–866 2001 .Yang, Z. Y., He, J. H., Lu, A. P., Hou, T. J. & Cao, D. S. Frequent hitters: nuisance artifacts in high-throughput screening. Drug Discov. Today 25 , 657–667 2020 .Fraser, J. S. & Murcko, M. A. Structure is beauty, but not always truth. Cell 187 , 517–520 2024 .Segler, M. H. S., Preuss, M. & Waller, M. P. Planning chemical syntheses with deep neural networks and symbolic AI. Nature 555 , 604–610 2018 .Genheden, S. et al. AiZynthFinder: a fast, robust and flexible open-source software for retrosynthetic planning. J. Cheminform. 12 , 70 2020 .Voinarovska, V., Kabeshov, M., Dudenko, D., Genheden, S. & Tetko, I. V. When yield prediction does not yield prediction: an overview of the current challenges. J. Chem. Inf. Model. 64 , 42–56 2024 .Gupta, A. How unfair is the coin? https://ankitg.me/blog/2025/01/06/unfair-coins.html https://ankitg.me/blog/2025/01/06/unfair-coins.html 2025 .Bilodeau, C. et al. Generative models for molecular discovery: recent advances and challenges. WIREs Comput. Mol. Sci. 12 , e1608 2022 .Gómez-Bombarelli, R. et al. Automatic chemical design using a data-driven continuous representation of molecules. ACS Cent. Sci. 4 , 268–276 2018 .Baillif, B., Cole, J., McCabe, P. & Bender, A. Deep generative models for 3D molecular structure. Curr. Opin. Struct. Biol. 80 , 102566 2023 .Cretu, M. et al. SynFlowNet: design of diverse and novel molecules with synthesis constraints. Preprint at https://arxiv.org/abs/2405.01155 https://arxiv.org/abs/2405.01155 2025 .Buttenschoen, M., Morris, G. M. & Deane, C. M. PoseBusters: AI-based docking methods fail to generate physically valid poses or generalise to novel sequences. Chem. Sci. 15 , 3130–3139 2023 .Handa, K. et al. On the difficulty of validating molecular generative models realistically: a case study on public and proprietary data. J. Cheminform. 15 , 112 2023 .Liu, H. et al. How good are current pocket-based 3D generative models?: the benchmark set and evaluation of protein pocket-based 3D molecular generative models. J. Chem. Inf. Model. 64 , 9260–9275 2024 .Nie, D. et al. Durian: a comprehensive benchmark for structure-Based 3D molecular generation. J. Chem. Inf. Model. 65 , 173–186 2025 .Baillif, B. et al. Benchmarking structure-based three-dimensional molecular generative models using GenBench3D: ligand conformation quality matters. Preprint at https://arxiv.org/abs/2407.04424 https://arxiv.org/abs/2407.04424 2024 .Renz, P. et al. On failure modes in molecule generation and optimization. Drug Discov. Today 32-33 , 55–63 2019 .Langevin, M., Vuilleumier, R. & Bianciotto, M. Explaining and avoiding failure modes in goal-directed generation of small molecules. J. Cheminform. 14 , 20 2022 .Senior, A. W. et al. Improved protein structure prediction using potentials from deep learning. Nature 577 , 706–710 2020 .Abramson, J. et al. Accurate structure prediction of biomolecular interactions with AlphaFold 3. Nature 630 , 493–500 2024 .Pereira, J. et al. High-accuracy protein structure prediction in CASP14. Proteins 89 , 1687–1699 2021 .Momin, A. A. et al. PYK2 senses calcium through a disordered dimerization and calmodulin-binding element. Commun. Biol. 5 , 800 2022 .Jänes, J. & Beltrao, P. Deep learning for protein structure prediction and design-progress and applications. Mol. Syst. Biol. 20 , 162–169 2024 .Watson, J. L. et al. De novo design of protein structure and function with RFdiffusion. Nature 620 , 1089–1100 2023 .Dauparas, J. et al. Robust deep learning–based protein sequence design using ProteinMPNN. Science 378 , 49–56 2022 .Goverde, C. A. et al. Computational design of soluble and functional membrane protein analogues. Nature 631 , 449–458 2024 .Li, Q., Vlachos, E. N. M. & Bryant, P. Design of linear and cyclic peptide binders from protein sequence information. Commun. Chem. 8 , 211 2025 .Lauko, A. et al. Computational design of serine hydrolases. Science 388 , eadu2454 2025 .Desautels, T. A. et al. Computationally restoring the potency of a clinical antibody against Omicron. Nature 629 , 878–885 2024 .Stark, H. et al. BoltzGen: toward universal binder design. Preprint at bioRxiv https://doi.org/10.1101/2025.11.20.689494 https://doi.org/10.1101/2025.11.20.689494 2025 .Chai Discovery Team et al. Zero-shot antibody design in a 24-well plate. Preprint at bioRxiv https://doi.org/10.1101/2025.07.05.663018 https://doi.org/10.1101/2025.07.05.663018 2025 .Boretti, A. Improving chimeric antigen receptor T-cell therapies by using artificial intelligence and internet of things technologies: a narrative review. Eur. J. Pharmacol. 974 , 176618 2024 .Bujak, J. et al. Creating an innovative artificial intelligence-based technology TCRact for designing and optimizing T cell receptors for use in cancer immunotherapies: protocol for an observational trial. JMIR Res. Protoc. 12 , e45872 2023 .Klontzas, M. E. et al. Machine learning and metabolomics predict mesenchymal stem cell oteogenic differentiation in 2D and 3D cultures. J. Funct. Biomater. 15 , 367 2024 .Bäckel, N. et al. Elaborating the potential of artificial intelligence in automated CAR-T cell manufacturing. Front. Mol. Med. 3 , 1250508 2023 .Duran, I. et al. Detection of senescence using machine learning algorithms based on nuclear features. Nat. Commun. 15 , 1–20 2024 .Bock, C. et al. High-content CRISPR screening. Nat. Rev. Methods Primers 2 , 8 2022 .Schmidhuber, J. 2011: DanNet triggers deep CNN revolution. https://people.idsia.ch/~juergen/DanNet-triggers-deep-CNN-revolution-2011.html https://people.idsia.ch/~juergen/DanNet-triggers-deep-CNN-revolution-2011.html 2021 .Rumelhart, D. E. & McClelland, J. L. in Parallel Distributed Processing: Explorations in the Microstructure of Cognition eds Feldman, J. A. et al. 318–362 MIT Press, 1987 .Archit, A. et al. Segment anything for microscopy. Nat. Methods 22 , 579–591 2025 .Kenyon-Dean, K. et al. ViTally consistent: scaling biological representation learning for cell microscopy. In Proc. 42nd International Conference on Machine Learning eds Singh, A. et al. Vol. 267, 29735–29752 PMLR, 2025 .Brown, T. B. et al. Language models are few-shot learners. In Proc. 34th International Conference on Neural Information Processing Systems eds Larochelle H. et al. article no. 159, 1877–1901 Curran Associates, Inc., 2020 .Asher, N. et al. Limits for learning with language models. In Proc. 12th Joint Conference on Lexical and Computational Semantics eds Palmer, A. & Camacho-Collados, J. 236–248 Association for Computational Linguistics, 2023 .Grisoni, F. Chemical language models for de novo drug design: challenges and opportunities. Curr. Opin. Struc. Biol. 97 , 102527 2023 .Bran, M. et al. Augmenting large language models with chemistry tools. Nat. Mach. Intell. 6 , 525–535 2024 .Chaves, J. M. Z. et al. Tx-LLM: a large language model for therapeutics. Preprint at https://arxiv.org/abs/2406.06316 https://arxiv.org/abs/2406.06316 2024 .Huynh, D. L. et al. AI agents in drug discovery: applications and case studies. Drug Discov. Today 31 , 104650 2026 .Hu, Q. et al. Machine learning to predict adverse drug events based on electronic health records: a systematic review and meta-analysis. J. Int. Med. Res. 52 , 3000605241302304 2024 .Singhal, P. et al. Opportunities and challenges for biomarker discovery using electronic health record data. Trends Mol. Med. 29 , 765–776 2023 .Moynihan, D. et al. Analysis and visualisation of electronic health records data to identify undiagnosed patients with rare genetic diseases. Sci. Rep. 14 , 5056 2024 .Yadav, P., Steinbach, M., Kumar, V. & Simon, G. Mining electronic health records EHRs : a survey. ACM Comput. Surv. 50 , 6 2018 .Sarwar, T. et al. The secondary use of electronic health records for data mining: data characteristics and challenges. ACM Comput. Surv . https://doi.org/10.1145/3490234 https://doi.org/10.1145/3490234 2023 .Gaber, F. et al. Evaluating large language model workflows in clinical decision support for triage and referral and diagnosis. npj Digit. Med. 8 , 263 2025 .Hager, P. et al. Evaluation and mitigation of the limitations of large language models in clinical decision-making. Nat. Med. 30 , 2613–2622 2024 .Ranji, S. R. Large language models — misdiagnosing diagnostic excellence? JAMA Netw. Open 7 , e2440901 2024 .Harrer, S., Shah, P., Antony, B. & Hu, J. Artificial intelligence for clinical trial design. Trends Pharmacol. Sci. 40 , 577–591 2019 .Deloitte Centre for Health Solutions. Intelligent clinical trials. Transforming through AI-enabled engagement. https://www2.deloitte.com/content/dam/insights/us/articles/22934 intelligent-clinical-trials/DI Intelligent-clinical-trials.pdf https://www2.deloitte.com/content/dam/insights/us/articles/22934 intelligent-clinical-trials/DI Intelligent-clinical-trials.pdf Deloitte University EMEA CVBA, 2024 .Razuvayevskaya, O. et al. Genetic factors associated with reasons for clinical trial stoppage. Nat. Genet. 56 , 1862–1867 2024 .Askin, S., Burkhalter, D., Calado, G. & El Dakrouni, S. Artificial intelligence applied to clinical trials: opportunities and challenges. Health Technol. 13 , 203–213 2023 .Ciray, F. & Doğan, T. Machine learning-based prediction of drug approvals using molecular, physicochemical, clinical trial, and patent-related features. Expert Opin. Drug Discov. 17 , 1425–1441 2022 .Aliper, A. et al. Prediction of clinical trials outcomes based on target choice and clinical trial design with multi-modal artificial intelligence. Clin. Pharmacol. Ther. 114 , 972–980 2023 .Zheng, W. et al. Multimodal clinical trial outcome prediction via large language models and mixture-of-experts. Findings Assoc. Comput. Linguistics EMNLP 2025 , 7503–7517 2025 .Proctor, W. R. et al. Utility of spherical human liver microtissues for prediction of clinical drug-induced liver injury. Arch. Toxicol. 91 , 2849–2863 2017 .Rudolf, A. F. et al. A comparison of protein kinases inhibitor screening methods using both enzymatic activity and binding affinity determination. PLoS ONE 10 , e98800 2014 .Huang, K. et al. Therapeutics data commons: machine learning datasets and tasks for drug discovery and development. Preprint at https://arxiv.org/abs/2102.09548 https://arxiv.org/abs/2102.09548 2021 .Goldstein, J. L. & Brown, M. S. The clinician-investigator: bewitched, bothered, and bewildered — but still beloved. J. Clin. Invest. 99 , 2803–2812 1997 .Hughes, J. P., Rees, S., Kalindjian, S. B. & Philpott, K. L. Principles of early drug discovery. Br. J. Pharmacol. 162 , 1239–1249 2011 .Hernández-Orozco, S., Kiani, N. A. & Zenil, H. Algorithmically probable mutations reproduce aspects of evolution, such as convergence rate, genetic memory and modularity. R. Soc. Open. Sci. 5 , 180399 2018 . Acknowledgements Research in the A.B. group is supported by the Center for Biotechnology, Khalifa University of Science and Technology KU-BTC , Khalifa University and PNRR grant no. 760066/23.05.2023, cod 83/15.11.2022. V. Curean is thanked for discussions about the manuscript. Author information Authors and Affiliations Corresponding author Ethics declarations Competing interests A.B. is a shareholder or options holder of or consultant with Healx, Pharmenable Therapeutics, Pangea Bio, Cresset Group, Evolvus, Harmonic Discovery, Ignota Labs, Turbine, VALID, Daimedex, EtioMap, Apheris, AstraZeneca and Novartis. M.C.T. has no competing interests to declare. J.W.S. is director and shareholder of JW Scannell Analytics, director and shareholder of Etheros Pharmaceuticals, holds equity options in Ochro Bio and has an advisory role with Hiro Capital. D.A.S. has no competing interests to declare. G.M.G. is a shareholder in AstraZeneca, Vertex and Nvidia. J.G.G. has no competing interests to declare. L.-L.P. has no competing Interests to declare. R.D.J. is CSO and cofounder of Powerhouse Biology and owns shares in the company. K.H. is an employee of Meiji Seika Pharma Co. and is a shareholder of Teijin Pharma. M.H. is an employee of Tanabe Pharma Corporation. S.S. is an employee of Human Chemical Company. M.M. has no competing interests to declare. M.F.S. is an employee and shareholder of biotx.ai. T.A. is CEO and cofounder of VALID and owns shares in the company. F.G. has no competing interests to declare. I.C.-C. has no competing interests to declare. Peer review Peer review information Nature Reviews Drug Discovery thanks T. Biancalani, W. Pitt and J. Stokes for their contribution to the peer review of this work. Additional information Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Related links Accurate predictions of novel biomolecular interactions with IsoDDE: https://zenodo.org/records/18606681 https://zenodo.org/records/18606681 ChEMBL: https://www.ebi.ac.uk/chembl/ https://www.ebi.ac.uk/chembl/ Drugs@FDA: https://www.accessdata.fda.gov/scripts/cder/daf/index.cfm https://www.accessdata.fda.gov/scripts/cder/daf/index.cfm Innovative Medicines Initiative: https://www.efpia.eu/about-medicines/development-of-medicines/public-private-partnerships https://www.efpia.eu/about-medicines/development-of-medicines/public-private-partnerships LIGAND-AI: https://ligand-ai.org https://ligand-ai.org OpenADMET: https://openadmet.ghost.io https://openadmet.ghost.io PubChem: https://pubchem.ncbi.nlm.nih.gov/ https://pubchem.ncbi.nlm.nih.gov/ Redefining drug discovery with AI: https://www.gene.com/stories/redefining-drug-discovery-with-ai https://www.gene.com/stories/redefining-drug-discovery-with-ai Tempus introduces Loop, an AI-powered target discovery and validation platform: https://www.tempus.com/news/tempus-introduces-loop-an-ai-powered-target-discovery-and-validation-platform https://www.tempus.com/news/tempus-introduces-loop-an-ai-powered-target-discovery-and-validation-platform Virtual Cell Pharmacology Initiative: https://datapoints.ginkgo.bio/updates/vcpi-blogpost https://datapoints.ginkgo.bio/updates/vcpi-blogpost Rights and permissions Springer Nature or its licensor e.g. a society or other partner holds exclusive rights to this article under a publishing agreement with the author s or other rightsholder s ; author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law. About this article Cite this article Bender, A., Thomas, M.C., Scannell, J.W. et al. Artificial intelligence in drug discovery — what it is, where we stand and the path forward. Nat Rev Drug Discov 2026 . https://doi.org/10.1038/s41573-026-01496-2 Accepted: Published: Version of record: DOI: https://doi.org/10.1038/s41573-026-01496-2