The role of AI in drug discovery: challenges, opportunities, and strategies
Open accessAbstract
Artificial intelligence (AI) has the potential to revolutionize the drug discovery process, offering improved efficiency, accuracy, and speed. However, the successful application of AI is dependent on the availability of high-quality data, the addressing of ethical concerns, and the recognition of the limitations of AI-based approaches. In this article, the benefits, challenges, and drawbacks of AI in this field are reviewed, and possible strategies and approaches for overcoming the present obstacles are proposed. The use of data augmentation, explainable AI, and the integration of AI with traditional experimental methods, as well as the potential advantages of AI in pharmaceutical research, are also discussed. Overall, this review highlights the potential of AI in drug discovery and provides insights into the challenges and opportunities for realizing its potential in this field. Note from the human authors: This article was created to test the ability of ChatGPT, a chatbot based on the GPT-3.5 language model, in terms of assisting human authors in writing review articles. The text generated by the AI following our instructions (see Supporting Information) was used as a starting point, and its ability to automatically generate content was evaluated. After conducting a thorough review, the human authors practically rewrote the manuscript, striving to maintain a balance between the original proposal and the scientific criteria. The advantages and limitations of using AI for this purpose are discussed in the last section.
AI snapshot
Written from the paper’s full text.
TL;DR
This review argues that AI can improve the efficiency, accuracy, and speed of drug discovery, but its success depends on high-quality data, ethical safeguards, and recognition of AI’s limits. It also reports an experiment in which ChatGPT-assisted drafting required substantial human rewriting and had only 6% correct references in the preliminary AI-generated text.
Key findings
- AI/ML can support prediction of drug efficacy and toxicity, drug–drug interactions, novel compound design, target identification, and case studies in cancer, Alzheimer’s, antibiotics, and COVID-19, though the review presents these as literature examples rather than new experimental results.
- The paper identifies major barriers: limited or low-quality training data, ethical concerns including bias and fairness, data privacy/security, and job-loss concerns from automation.
- Proposed strategies include data augmentation, explainable AI (XAI), and integration of AI with traditional experimental methods and human expertise.
- In the writing experiment, the preliminary ChatGPT text had only 6% correct references; similarity to the final manuscript was identical 4.3%, minor changes 13.3%, and related meaning 16.3%. The authors conclude ChatGPT is not useful for reliable scientific texts without substantial human intervention.
Methods
- Narrative review on AI in drug discovery, generated with assistance from ChatGPT (GPT-3.5), with human authors providing topic, section count, prompts, and instructions.
- Human authors edited AI-generated text, corrected/enriched content, avoided repetitions and inconsistencies, and revised all AI-suggested references through iterative revisions.
- The paper evaluated the AI-assisted process using similarity percentages between preliminary and final text and the percentage of correct references in the preliminary ChatGPT output.
Limitations
- Stated: AI-based approaches require large volumes of suitable data; accessible data may be limited, low quality, or inconsistent, affecting accuracy and reliability.
- Stated: AI cannot substitute for traditional experimental methods or human researchers’ expertise; predictions must be validated and interpreted by humans.
- Stated: ChatGPT lacked ability to evaluate veracity/reliability, was trained in 2021 and not updated, and produced mostly incorrect references; the authors also note reviewers may struggle to distinguish AI-written from human-written articles. Evident from the text: this is a narrative review without a systematic search strategy, and its drug-discovery examples are cited from other studies rather than independently validated here.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
0 of 67 cited sources checked · 0 supported · 0 flagged
- Interpretationcitation
AI-based approaches can enable the rapid and efficient design of novel compounds with desirable properties and activities.
“AI-based approaches, on the other hand, can enable the rapid and efficient design of novel compounds with desirable properties and activities.”
Quote in the paperSubstantial claim - Premisecitation
Combining ML and MD methodologies is being explored to take advantage of synergies between them.
“The technique of combining these methodologies is being explored to take advantage of the synergies between them [20].”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [20] Bai, 2022, Application advances of deep learning methods for de novo drug design and molecular dynamics simulation, Wiley Interdiscip. Rev. Comput. Mol. Sci. source
Not checked in the free run.
- Interpretationreasoning
The use of AI in pharmaceutical research offers the ability to improve the prediction of the efficacy and toxicity of potential drug compounds.
“The previous examples from the literature demonstrate that the use of AI in pharmaceutical research offers the ability to improve the prediction of the efficacy and toxicity of potential drug compounds.”
Quote in the paperSubstantial claim - Premisecitation
A deep learning algorithm trained on a dataset of known drug compounds and their biological activity was able to predict the activity of novel compounds with high accuracy.
“For instance, a DL algorithm has recently been trained using a dataset of known drug compounds, along with their corresponding biological activity [11]. The algorithm was then able to predict the activity of novel compounds with high accuracy.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [11] Hansen, 2015, Machine learning predictions of molecular properties: Accurate many-body potentials and nonlocality in chemical space, J. Phys. Chem. Lett. source
Not checked in the free run.
- Premisereasoning
AI-based approaches can identify drug-drug interactions by analyzing large datasets of known drug interactions and recognizing patterns and trends.
“This issue can be identified by AI-based approaches by analyzing large datasets of known drug interactions and recognizing the patterns and trends.”
Quote in the paperSubstantial claim
- Premisereasoning
The use of interpretable machine learning and deep learning methods is contributing to de novo drug design.
“The use of interpretable machine learning (IML) and DL methods is also contributing to this effort.”
Quote in the paperNot a substantial claim - Premisecitation
AI algorithms including supervised and unsupervised learning, reinforcement, and evolutionary or rule-based algorithms can potentially contribute to solving problems of current medicinal chemistry methods.
“Different algorithms based on AI, including supervised and unsupervised learning methods, reinforcement, and evolutionary or rule-based algorithms, can potentially contribute to solving these problems.”
Quote in the paperSubstantial claim0 of 3 sources checked- cites [9] Chen, R., Liu, X., Jin, S., Lin, J., and Liu, J. (2018). Machine learning for drug-target interaction prediction. Molecules, 23. source
Not checked in the free run.
- cites [10] Wei, 2018, Automatic Chemical Design Using a Data-Driven Continuous Representation of Molecules, ACS Central Sci. source
Not checked in the free run.
- cites [11] Hansen, 2015, Machine learning predictions of molecular properties: Accurate many-body potentials and nonlocality in chemical space, J. Phys. Chem. Lett. source
Not checked in the free run.
- Premisecitation
A DL algorithm trained on a large dataset of known cancer-related compounds and their biological activity produced novel compounds with high potential for future cancer treatment.
“These authors trained a DL algorithm on a large dataset of known cancer-related compounds and their corresponding biological activity. As an output, novel compounds with high potential for future cancer treatment were obtained, demonstrating the ability of this method to discover new therapeutic candidates.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [21] Gupta, 2021, Artificial intelligence to deep learning: Machine intelligence approach for drug discovery, Mol. Divers. source
Not checked in the free run.
- Premisecitation
ML techniques and molecular dynamics simulations are currently being used in de novo drug design to improve efficiency and accuracy.
“ML techniques and molecular dynamics (MD) simulations are currently being used in the field of de novo drug design to improve efficiency and accuracy.”
Quote in the paperSubstantial claim0 of 1 sources checked- cites [20] Bai, 2022, Application advances of deep learning methods for de novo drug design and molecular dynamics simulation, Wiley Interdiscip. Rev. Comput. Mol. Sci. source
Not checked in the free run.
- Premisereasoning
ML algorithms can identify patterns and trends that may not be apparent to human researchers based on analysis of large amounts of information.
“Based on the analysis of a large amount of information, ML algorithms can identify patterns and trends that may not be apparent to human researchers.”
Quote in the paperSubstantial claim - Premisecitation
A deep learning algorithm trained on a dataset of known drug compounds and their corresponding properties proposed new therapeutic molecules with desirable characteristics such as solubility and activity.
“For example, a deep learning (DL) algorithm has recently been trained on a dataset of known drug compounds and their corresponding properties, to propose new therapeutic molecules [10] with desirable characteristics such as solubility and activity, demonstrating the potential of these methods for the rapid and efficient design of new drug candidates.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [10] Wei, 2018, Automatic Chemical Design Using a Data-Driven Continuous Representation of Molecules, ACS Central Sci. source
Not checked in the free run.
- Premisecitation
Deep learning has been successfully used to predict the efficacy of drug compounds with high accuracy.
“The successful use of deep learning (DL) to predict the efficacy of drug compounds with high accuracy has been described recently by the authors of [5].”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [5] Zhuang, D., and Ibrahim, A.K. (2021). Deep learning for drug discovery: A study of identifying high efficacy drug compounds using a cascade transfer learning approach. Appl. Sci., 11. source
Not checked in the free run.
- Premisecitation
Currently, medicinal chemistry methods rely heavily on a hit-and-miss approach and large-scale testing techniques.
“Currently, medicinal chemistry methods rely heavily on a hit-and-miss approach and large-scale testing techniques [8].”
Quote in the paperSubstantial claim0 of 1 sources checked- cites [8] Wess, 2001, Medicinal Chemistry: Challenges and Opportunities, Angew. Chem. Int. Ed. source
Not checked in the free run.
- Premisecitation
ML trained on large databases of known toxic and non-toxic compounds has made significant contributions to prevent the toxicity of potential drug compounds.
“Significant contributions to prevent the toxicity of potential drug compounds, employing intensive training using large databases of known toxic and non-toxic compounds for ML, have also been published [17].”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [17] Solana, 2021, Toxicity prediction based on artificial intelligence: A multidisciplinary overview, Wiley Interdiscip. Rev. Comput. Mol. Sci. source
Not checked in the free run.
- Premisecitation
AI-based methods have been able to predict the toxicity of drug candidates.
“AI-based methods have also been able to predict the toxicity of drug candidates [6].”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [6] Pu, L., Naderi, M., Liu, T., Wu, H.C., Mukhopadhyay, S., and Brylinski, M. (2019). EToxPred: A machine learning-based approach to estimate the toxicity of drug candidates. BMC Pharmacol. Toxicol., 20. source
Not checked in the free run.
- Premisecitation
AI-based algorithms can be employed to identify new targets for drug development, such as specific proteins or genetic pathways involved in diseases.
“Furthermore, AI-based algorithms can also be employed to identify new targets for drug development, such as the specific proteins or genetic pathways involved in diseases [14].”
Quote in the paperSubstantial claim0 of 1 sources checked- cites [14] You, 2019, Predicting drug-target interaction network using deep learning model, Comput. Biol. Chem. source
Not checked in the free run.
- Premisecitation
The efficacy and toxicity of new drug compounds can be predicted using AI approaches with greater accuracy and efficiency than traditional methods.
“For instance, the efficacy and toxicity of new drug compounds can be predicted using these approaches, with greater accuracy and efficiency than when using traditional methods [12,13].”
Quote in the paperSubstantial claim0 of 2 sources checked - Interpretationreasoning
ML can enable the proposal of new bioactive compounds with minimum side effects in a much faster process than classical protocols.
“This can enable the proposal of new bioactive compounds with minimum side effects in a much faster process than when using classical protocols.”
Quote in the paperSubstantial claim - Premisecitation
AlphaFold is a software platform that uses protein sequence data and AI to predict proteins' corresponding three-dimensional structures.
“It is a powerful algorithm that uses protein sequence data and AI to predict the proteins’ corresponding three-dimensional structures.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [19] Nussinov, 2022, AlphaFold, Artificial Intelligence (AI), and Allostery, J. Phys. Chem. B source
Not checked in the free run.
- Premisecitation
AI has been successfully used to identify novel compounds for the treatment of cancer.
“For example, the successful use of AI to identify novel compounds for the treatment of cancer has recently been reported by Gupta, R., et al. [21].”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [21] Gupta, 2021, Artificial intelligence to deep learning: Machine intelligence approach for drug discovery, Mol. Divers. source
Not checked in the free run.
- Premisecitation
An ML algorithm has been used to accurately predict the interactions of novel drug pairs.
“This has been recently addressed by an ML algorithm used to accurately predict the interactions of novel drug pairs [18].”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [18] Jang, 2022, Machine learning-based quantitative prediction of drug exposure in drug-drug interactions using drug label information, npj Digit. Med. source
Not checked in the free run.
- Premisecitation
AI techniques such as machine learning and natural language processing offer the potential to accelerate and improve drug discovery by enabling more efficient and accurate analysis of large amounts of data.
“However, AI techniques such as machine learning (ML) and natural language processing offer the potential to accelerate and improve this process by enabling more efficient and accurate analysis of large amounts of data [4].”
Quote in the paperSubstantial claim0 of 1 sources checked- cites [4] Xu, 2021, Artificial intelligence: A powerful paradigm for scientific research, Innovation
Not checked in the free run.
- Interpretationcitation
The potential of AI in drug discovery has been demonstrated in several case studies.
“The potential of AI in the context of drug discovery has been demonstrated in several case studies.”
Quote in the paperNot a substantial claim - Premisecitation
Ethical considerations are another challenge because AI-based approaches may raise concerns about fairness and bias.
“Another challenge is presented by ethical considerations [42] since AI-based approaches may raise concerns about fairness and bias (see next section) [43].”
Quote in the paperNot a substantial claim0 of 2 sources checked- cites [42] Basu, 2020, The ethics of machine learning in medical sciences: Where do we stand today?, Indian J. Dermatol. source
Not checked in the free run.
- cites [43] Kleinberg, J. (2018, January 18–22). Inherent Trade-Offs in Algorithmic Fairness. Proceedings of the Abstracts of the 2018 ACM International Conference on Measurement and Modeling of Computer Systems, Irvine, CA, USA. source
Not checked in the free run.
- Premisecitation
In some cases, AI-powered approaches have identified promising drug candidates for COVID-19 in a fraction of the time that traditional methods would take.
“In some cases, these AI-powered approaches have been able to identify promising drug candidates in a fraction of the time that it would take when using traditional methods [26,27,28,29,30,31].”
Quote in the paperSubstantial claim0 of 6 sources checked- cites [26] Lv, 2021, Application of artificial intelligence and machine learning for COVID-19 drug discovery and vaccine design, Brief. Bioinform. source
Not checked in the free run.
- cites [27] Monteleone, 2022, Fighting COVID-19 with Artificial Intelligence, Methods in Molecular Biology source
Not checked in the free run.
- cites [28] Zhou, 2020, Artificial intelligence in COVID-19 drug repurposing, Lancet Digit. Health source
Not checked in the free run.
- cites [29] Verma, N., Qu, X., Trozzi, F., Elsaied, M., Karki, N., Tao, Y., Zoltowski, B., Larson, E.C., and Kraka, E. (2021). Predicting potential Sars-Cov-2 drugs-in depth drug database screening using deep neural network framework ssnet, classical virtual screening and docking. Int. J. Mol. Sci., 22. source
Not checked in the free run.
- cites [30] Bung, 2021, De novo design of new chemical entities for SARS-CoV-2 using artificial intelligence, Future Med. Chem. source
Not checked in the free run.
- cites [31] Floresta, G., Zagni, C., Gentile, D., Patamia, V., and Rescifina, A. (2022). Artificial Intelligence Technologies for COVID-19 De Novo Drug Design. Int. J. Mol. Sci., 23. source
Not checked in the free run.
- Premisecitation
Data augmentation can increase the quantity and diversity of data available for training ML algorithms, improving the accuracy and reliability of results.
“One approach is the use of data augmentation [46], which involves the generation of synthetic data to supplement existing datasets. This can increase the quantity and diversity of the data available for training ML algorithms, improving the accuracy and reliability of the results [47].”
Quote in the paperNot a substantial claim0 of 2 sources checked- cites [46] Kusam, L., Mayank, D., and Nishanth, K.N. (2019, January 8). Data Augmentation Using Generative Adversarial Network. Proceedings of the 2nd International Conference on Advanced Computing and Software Engineering (ICACSE) 2019, Sultanpur, India.
Not checked in the free run.
- cites [47] Taylor, L., and Nitschke, G. (2018, January 18–21). Improving Deep Learning with Generic Data Augmentation. Proceedings of the 2018 IEEE Symposium Series on Computational Intelligence, SSCI 2018, Piscataway, NJ, USA. source
Not checked in the free run.
- Premisecitation
AI-based approaches typically require a large volume of information for training purposes.
“AI-based approaches typically require a large volume of information for training purposes [41].”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [41] Tsuji, 2021, Artificial intelligence-based computational framework for drug-target prioritization and inference of novel repositionable drugs for Alzheimer’s disease, Alzheimer Res. Ther. source
Not checked in the free run.
- Premisecitation
AI has been successfully applied in the discovery of new antibiotics.
“AI has also been successfully applied in the discovery of new antibiotics [24].”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [24] Melo, 2021, Accelerating antibiotic discovery through artificial intelligence, Commun. Biol. source
Not checked in the free run.
- Premisecitation
One key challenge of using AI in drug discovery is the availability of suitable data.
“One of the key challenges is the availability of suitable data [40].”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [40] Vamathevan, 2019, Applications of machine learning in drug discovery and development, Nat. Rev. Drug Discov. source
Not checked in the free run.
- Premisecitation
ML has been used to identify small-molecule inhibitors of the protein MEK.
“The use of ML to identify small-molecule inhibitors of the protein MEK [22] has recently been described.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [22] Zhu, 2021, Prediction of drug efficacy from transcriptional profiles with deep learning, Nat. Biotechnol. source
Not checked in the free run.
- Interpretationcitation
Integration of AI with traditional experimental methods can enhance the drug discovery process.
“However, the integration of AI with traditional experimental methods can also enhance the drug discovery process [3].”
Quote in the paperSubstantial claim0 of 1 sources checked- cites [3] Paul, 2021, Artificial intelligence in drug discovery and development, Drug Discov. Today source
Not checked in the free run.
- Interpretationcitation
Combining the predictive power of AI with the expertise and experience of human researchers can optimize the drug discovery process and accelerate the development of new medications.
“By combining the predictive power of AI with the expertise and experience of human researchers [53], it is possible to optimize the drug discovery process and accelerate the development of new medications [54].”
Quote in the paperSubstantial claim0 of 2 sources checked- cites [53] Jarrahi, 2018, Artificial intelligence and the future of work: Human-AI symbiosis in organizational decision making, Bus. Horiz. source
Not checked in the free run.
- cites [54] Wang, 2019, Artificial intelligence facilitates drug design in the big data era, Chemom. Intell. Lab. Syst. source
Not checked in the free run.
- Premisecitation
ML has been used to identify novel inhibitors of beta-secretase (BACE1), a protein involved in Alzheimer's disease.
“Another example is the identification of novel inhibitors of beta-secretase (BACE1), a protein involved in the development of Alzheimer’s disease [23] by using an ML algorithm.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [23] Dhamodharan, 2022, Machine learning models for predicting the activity of AChE and BACE1 dual inhibitors for the treatment of Alzheimer’s disease, Mol. Divers. source
Not checked in the free run.
- Premisecitation
If the data used to train an ML algorithm are biased or unrepresentative, the resulting predictions may be inaccurate or unfair.
“For example, if the data used to train an ML algorithm are biased or unrepresentative, the resulting predictions may be inaccurate or unfair [44].”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [44] Silvia, 2020, When Worlds Collide: Protecting Physical World Interests Against Virtual World Malfeasance, Michigan Technol. Law Rev. source
Not checked in the free run.
- Premisecitation
An ML algorithm was able to identify novel inhibitors for the protein MEK.
“The ML algorithm was able to identify novel inhibitors for this protein.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [22] Zhu, 2021, Prediction of drug efficacy from transcriptional profiles with deep learning, Nat. Biotechnol. source
Not checked in the free run.
- Premisecitation
ML algorithms have been used to analyze large datasets of potential compounds and identify those with the most potential for treating COVID-19.
“ML algorithms have been used to analyze large datasets of potential compounds and identify those with the most potential for treating the virus.”
Quote in the paperSubstantial claim0 of 6 sources checked- cites [26] Lv, 2021, Application of artificial intelligence and machine learning for COVID-19 drug discovery and vaccine design, Brief. Bioinform. source
Not checked in the free run.
- cites [27] Monteleone, 2022, Fighting COVID-19 with Artificial Intelligence, Methods in Molecular Biology source
Not checked in the free run.
- cites [28] Zhou, 2020, Artificial intelligence in COVID-19 drug repurposing, Lancet Digit. Health source
Not checked in the free run.
- cites [29] Verma, N., Qu, X., Trozzi, F., Elsaied, M., Karki, N., Tao, Y., Zoltowski, B., Larson, E.C., and Kraka, E. (2021). Predicting potential Sars-Cov-2 drugs-in depth drug database screening using deep neural network framework ssnet, classical virtual screening and docking. Int. J. Mol. Sci., 22. source
Not checked in the free run.
- cites [30] Bung, 2021, De novo design of new chemical entities for SARS-CoV-2 using artificial intelligence, Future Med. Chem. source
Not checked in the free run.
- cites [31] Floresta, G., Zagni, C., Gentile, D., Patamia, V., and Rescifina, A. (2022). Artificial Intelligence Technologies for COVID-19 De Novo Drug Design. Int. J. Mol. Sci., 23. source
Not checked in the free run.
- Interpretationown data
ChatGPT lacks the knowledge and expertise necessary to accurately and adequately convey complex scientific concepts and information.
“The program lacks the knowledge and expertise necessary to accurately and adequately convey complex scientific concepts and information.”
Quote in the paperNot a substantial claim - Interpretationown data
ChatGPT is not a useful tool for writing reliable scientific texts without substantial human intervention.
“As a result of this experiment, we can state that ChatGPT is not a useful tool for writing reliable scientific texts without substantial human intervention.”
Quote in the paperSubstantial claim
- Classified as probably not a substantial claim. 29
- May combine two claims. 22
Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.
FindingcitationThis paper was generated with the assistance of ChatGPT, a chatbot based on the GPT-3.5 language model, trained with a large corpus of text via OpenAI, which at that time did not have connection to the Internet.
“This paper was generated with the assistance of ChatGPT, a chatbot based on the GPT-3.5 language model, trained with a large corpus of text via OpenAI [1], which at that time did not have connection to the Internet.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [1] (2022, December 06). OpenAI. ChatGPT (Dec 6 Version) [Large Language Model]. Available online: https://chat.openai.com/chat.
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.
FindingcitationThe total percentage of similarity between the preliminary text obtained directly from ChatGPT and the current version of the manuscript is: identical 4.3%, minor changes 13.3%, and related meaning 16.3%.
“The total percentage of similarity between the preliminary text, obtained directly from ChatGPT, and the current version of the manuscript is: identical 4.3%, minor changes 13.3%, and related meaning 16.3% [2].”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [2] (2022, December 06). Plagiarism Detector Software|Anti-Plagiarism Tools|Copyleaks. Available online: https://copyleaks.com.
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
Findingnone givenThe percentage of correct references in the preliminary text obtained directly from ChatGPT was just 6%.
“The percentage of correct references in the preliminary text, obtained directly from ChatGPT, was just 6%.”
Quote in the paperNot a substantial claimClassified as probably not a substantial claim. May combine two claims.
Findingown dataIn the authors' experiment, revision of AI-generated text required major edits and corrections, including replacement of nearly all references because those provided by ChatGPT were clearly incorrect.
“Our revision of the text, which was generated by the AI following our instructions, required the application of major edits and corrections, including the replacement of nearly all the references since those provided by the software were clearly incorrect.”
Quote in the paperSubstantial claimProbably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.
Findingnone givenChatGPT was trained in 2021 and does not work with updated information.
“Another important problem of the employed AI-based tool is that it was trained in 2021 and so does not work with updated information.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.
FindingreasoningChatGPT has advantages as an AI assistant to write scientific papers, including capacity to generate and optimize text quickly, and ability to help with organization of information or connecting ideas.
“As an AI assistant to write scientific papers, ChatGPT has several advantages, including its capacity to generate and optimize text quickly, as well as its ability to help users with several tasks, including the organization of information or even connecting ideas in some cases.”
Quote in the paperNot a substantial claimClassified as probably not a substantial claim. May combine two claims.
InterpretationreasoningThe successful application of AI in drug discovery is dependent on the availability of high-quality data, the addressing of ethical concerns, and the recognition of the limitations of AI-based approaches.
“However, the successful application of AI is dependent on the availability of high-quality data, the addressing of ethical concerns, and the recognition of the limitations of AI-based approaches.”
- Also stated in: § 10. Conclusions and Summary of the Potential of AI for Revolutionizing Drug Discovery · plain
“However, the successful application of AI in drug discovery is dependent on the availability of high-quality data, the addressing of ethical concerns, and the recognition of the limitations of AI-based approaches.”
Quote in the paperSubstantial claim- Also stated in: § 10. Conclusions and Summary of the Potential of AI for Revolutionizing Drug Discovery · plain
Probably not a substantial claim: may not pass S4 (based).
InterpretationreasoningAI has the potential to revolutionize the drug discovery process, offering improved efficiency, accuracy, and speed.
“Artificial intelligence (AI) has the potential to revolutionize the drug discovery process, offering improved efficiency, accuracy, and speed.”
- Also stated in: § 10. Conclusions and Summary of the Potential of AI for Revolutionizing Drug Discovery · plain
“In conclusion, AI has the potential to revolutionize the drug discovery process, offering improved efficiency and accuracy, accelerated drug development, and the capacity for the development of more effective and personalized treatments (Figure 1).”
Quote in the paperNot a substantial claim- Also stated in: § 10. Conclusions and Summary of the Potential of AI for Revolutionizing Drug Discovery · plain
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationThe use of AI in medicinal chemistry has gained significant attention in recent years as a potential means of revolutionizing the pharmaceutical industry.
“The use of artificial intelligence (AI) in medicinal chemistry has gained significant attention in recent years as a potential means of revolutionizing the pharmaceutical industry [3].”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [3] Paul, 2021, Artificial intelligence in drug discovery and development, Drug Discov. Today source
Not checked in the free run.
Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).
InterpretationreasoningAlphaFold represents a significant step forward in the use of AI in structural biology and life sciences.
“AlphaFold represents a significant step forward in the use of AI in structural biology and life sciences in general.”
Quote in the paperNot a substantial claimMay combine two claims.
InterpretationreasoningAI-based approaches have the ability to improve the efficiency and accuracy of drug discovery processes and can lead to the development of more effective medications.
“AI-based approaches, on the other hand, have the ability to improve the efficiency and accuracy of drug discovery processes and can lead to the development of more effective medications.”
Quote in the paperSubstantial claimMay combine two claims.
PremisecitationTraditional methods of pharmaceutical research are limited by their reliance on trial-and-error experimentation and their inability to accurately predict the behavior of new potential bioactive compounds.
“Thus, while traditional methods of pharmaceutical research have been relatively successful in the past, they are limited by their reliance on trial-and-error experimentation and their inability to accurately predict the behavior of new potential bioactive compounds [16].”
Quote in the paperSubstantial claim0 of 1 sources checked- cites [16] Bannigan, 2021, Machine learning directed drug formulation development, Adv. Drug Deliv. Rev. source
Not checked in the free run.
May combine two claims.
PremisecitationCurrent medicinal chemistry methods can be slow, costly, and often yield results with low accuracy.
“However, these methods can be slow, costly, and often yield results with low accuracy [6].”
Quote in the paperSubstantial claim0 of 1 sources checked- cites [6] Pu, L., Naderi, M., Liu, T., Wu, H.C., Mukhopadhyay, S., and Brylinski, M. (2019). EToxPred: A machine learning-based approach to estimate the toxicity of drug candidates. BMC Pharmacol. Toxicol., 20. source
Not checked in the free run.
May combine two claims.
InterpretationreasoningAI can identify possible drug-drug interactions in personalized medicine, enabling custom-made treatment plans that minimize the risk of adverse reactions.
“The role of AI to identify possible drug–drug interactions in the context of personalized medicine is also relevant, enabling the development of custom-made treatment plans that minimize the risk of adverse reactions.”
Quote in the paperSubstantial claimMay combine two claims.
InterpretationreasoningAI can enable the development of more effective and safer medications and accelerate the drug discovery process.
“This can enable the development of more effective and safer medications and accelerate the drug discovery process.”
Quote in the paperSubstantial claimMay combine two claims.
PremisecitationCurrent medicinal chemistry methods can be limited by the availability of suitable test compounds and the difficulty of accurately predicting their behavior in the body.
“In addition, they can be limited by the availability of suitable test compounds and the difficulty of accurately predicting their behavior in the body [9].”
Quote in the paperSubstantial claim0 of 1 sources checked- cites [9] Chen, R., Liu, X., Jin, S., Lin, J., and Liu, J. (2018). Machine learning for drug-target interaction prediction. Molecules, 23. source
Not checked in the free run.
Classified as probably not a substantial claim.
InterpretationreasoningCollaboration between AI researchers and pharmaceutical scientists can help improve the accuracy and efficiency of clinical trials by using AI algorithms to analyze trial data to identify trends and potential adverse effects.
“This collaboration can also help improve the accuracy and efficiency of clinical trials, as AI algorithms can be used to analyze the data collected during these trials to identify trends and the potential adverse effects of the drugs being tested.”
Quote in the paperSubstantial claimMay combine two claims.
InterpretationcitationAI-based methods can accelerate the drug discovery process and enable the development of more effective medications.
“Many more examples are available [3,32,33,34,35,36,37], showing that AI-based methods can accelerate the drug discovery process and enable the development of more effective medications.”
Quote in the paperSubstantial claim0 of 7 sources checked- cites [3] Paul, 2021, Artificial intelligence in drug discovery and development, Drug Discov. Today source
Not checked in the free run.
- cites [32] Vatansever, 2021, Artificial intelligence and machine learning-aided drug discovery in central nervous system diseases: State-of-the-arts and future directions, Med. Res. Rev. source
Not checked in the free run.
- cites [33] Farghali, 2021, The Potential Applications of Artificial Intelligence in Drug Discovery and Development, Physiol. Res. source
Not checked in the free run.
- cites [34] Ganesh, G.S., Kolusu, A.S., Prasad, K., Samudrala, P.K., and Nemmani, K.V.S. (2022). Advancing health care via artificial intelligence: From concept to clinic. Eur. J. Pharmacol., 934. source
Not checked in the free run.
- cites [35] Koromina, 2019, Rethinking Drug Repositioning and Development with Artificial Intelligence, Machine Learning, and Omics, OMICS A J. Integr. Biol. source
Not checked in the free run.
- cites [36] Mak, 2019, Artificial intelligence in drug development: Present status and future prospects, Drug Discov. Today source
Not checked in the free run.
- cites [37] Fleming, 2018, How artificial intelligence is changing drug discovery spotlight, Nature source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).
Premisenone givenAI in the pharmaceutical industry raises concerns about job losses due to automation.
“The use of AI in the pharmaceutical industry also raises concerns about job losses due to automation.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing).
PremisereasoningAI in the pharmaceutical industry raises questions about data privacy and security.
“Additionally, the use of AI in the pharmaceutical industry raises questions about data privacy and security.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing).
InterpretationreasoningAs AI systems rely on large amounts of data, there is a risk that sensitive personal information could be accessed or misused.
“As AI systems rely on large amounts of data to function, there is a risk that sensitive personal information could be accessed or misused.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing).
InterpretationreasoningUsing AI algorithms to analyze data from large populations can identify trends and patterns that help predict the effectiveness of potential drug candidates for specific patient populations and tailor treatments.
“By using AI algorithms to analyze data from large populations, they can be used to identify trends and patterns that can help predict the effectiveness of potential drug candidates for specific patient populations, which can help tailor treatments to the needs of individual patients.”
Quote in the paperNot a substantial claimMay combine two claims.
PremisecitationThe amount of accessible data may be limited, or data may be of low quality or inconsistent, which can affect the accuracy and reliability of AI results.
“In many cases, the amount of data that is accessible may be limited, or the data may be of low quality or inconsistent, which can affect the accuracy and reliability of the results [10].”
Quote in the paperSubstantial claim0 of 1 sources checked- cites [10] Wei, 2018, Automatic Chemical Design Using a Data-Driven Continuous Representation of Molecules, ACS Central Sci. source
Not checked in the free run.
May combine two claims.
InterpretationcitationAI can only provide predictions based on the data available, and the results must be validated and interpreted by human researchers.
“AI can only provide predictions based on the data available, and the results must then be validated and interpreted by human researchers [52].”
Quote in the paperSubstantial claim0 of 1 sources checked- cites [52] Gilpin, L.H., Bau, D., Yuan, B.Z., Bajwa, A., Specter, M., and Kagal, L. (2018, January 1–3). Explaining explanations: An overview of interpretability of machine learning. Proceedings of the 2018 IEEE 5th International Conference on Data Science and Advanced Analytics, DSAA Turin, Itali. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.
PremisecitationMany new companies are arising around AI in drug discovery and their impact is expected to be significant in the short term.
“Many new companies are currently arising around this area of research and their impact is expected to be significant in the short term [39].”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [39] (2022, December 06). 11 Companies Using Pharma AI to Stimulate Growth in the Industry. Available online: https://www.p360.com/data360/11-companies-using-pharma-ai-to-stimulate-growth-in-the-industry-1/.
Not checked in the free run.
Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).
InterpretationreasoningCollaboration between AI researchers and pharmaceutical scientists can help improve the accessibility and affordability of healthcare.
“Furthermore, collaboration between AI researchers and pharmaceutical scientists can also help to improve the accessibility and affordability of healthcare.”
Quote in the paperNot a substantial claimMay combine two claims.
PremisecitationExplainable AI methods can help address concerns about bias and fairness in AI-based approaches and provide a better understanding of the underlying mechanisms and assumptions behind predictions.
“This can help to address concerns about bias and fairness in AI-based approaches [43] and provide a better understanding of the underlying mechanisms and assumptions behind the predictions [49].”
Quote in the paperNot a substantial claim0 of 2 sources checked- cites [43] Kleinberg, J. (2018, January 18–22). Inherent Trade-Offs in Algorithmic Fairness. Proceedings of the Abstracts of the 2018 ACM International Conference on Measurement and Modeling of Computer Systems, Irvine, CA, USA. source
Not checked in the free run.
- cites [49] Arrieta, 2020, Explainable Artificial Intelligence (XAI): Concepts, taxonomies, opportunities and challenges toward responsible AI, Inf. Fusion source
Not checked in the free run.
Classified as probably not a substantial claim.
InterpretationreasoningCollaboration between AI researchers and pharmaceutical scientists can help create powerful algorithms and machine-learning models intended to predict the efficacy of potential drug candidates and speed up the drug discovery process.
“By combining their expertise and knowledge, they can create powerful algorithms and machine-learning models intended to predict the efficacy of potential drug candidates and speed up the drug discovery process.”
Quote in the paperSubstantial claimMay combine two claims.
InterpretationcitationCurrent AI-based approaches are not a substitute for traditional experimental methods and cannot replace the expertise and experience of human researchers.
“Current AI-based approaches are not a substitute for traditional experimental methods, and they cannot replace the expertise and experience of human researchers [50,51].”
Quote in the paperSubstantial claim0 of 2 sources checked- cites [50] Grebner, 2021, Application of Deep Neural Network Models in Drug Discovery Programs, ChemMedChem source
Not checked in the free run.
- cites [51] Schraagen, J.M., and van Diggelen, J. (2021). Expertise at Work, Palgrave Macmillan.
Not checked in the free run.
Classified as probably not a substantial claim.
InterpretationreasoningRecent developments in AI, including data augmentation, explainable AI, and integration with traditional experimental methods, offer promising strategies for overcoming challenges and limitations of AI in drug discovery.
“Recent developments in AI, including the use of data augmentation, explainable AI, and the integration of AI with traditional experimental methods, offer promising strategies for overcoming the challenges and limitations of AI in the context of drug discovery.”
Quote in the paperSubstantial claimMay combine two claims.
PremisecitationA pioneering ML approach identified powerful types of antibiotic from a pool of more than 100 million molecules, including one that works against a wide range of bacteria, such as tuberculosis and untreatable bacterial strains.
“A pioneering ML approach has identified powerful types of antibiotic from a pool of more than 100 million molecules, including one that works against a wide range of bacteria, such as tuberculosis and untreatable bacterial strains [25].”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [25] Marchant, J. (Nature, 2020). Powerful antibiotics discovered using AI, Nature, Online ahead of print. source
Not checked in the free run.
Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based). May combine two claims.
InterpretationreasoningGrowing levels of interest and attention from researchers, pharmaceutical companies, and regulatory agencies, combined with the potential benefits of AI, make this an exciting and promising area of research, with the potential to transform the drug discovery process.
“The growing levels of interest and attention from researchers, pharmaceutical companies, and regulatory agencies, combined with the potential benefits of AI, make this an exciting and promising area of research, with the potential to transform the drug discovery process.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationMerck and Numerate collaborated to develop AI-based approaches for medicinal chemistry.
“An illustrative example is the collaboration between the pharmaceutical company Merck and the AI company Numerate to develop AI-based approaches for medicinal chemistry [38].”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [38] (2022, December 06). Available online: https://www.fiercebiotech.com/biotech/numerate-forms-drug-discovery-collaboration-merck-to-utilize-numerate-s-silico-drug-design.
Not checked in the free run.
Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).
Interpretationown dataThe language and style used by ChatGPT may not be suitable for academic writing.
“Additionally, the language and style used by ChatGPT may not be suitable for academic writing.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing). May combine two claims.
InterpretationreasoningMost of these problems could be solved relatively quickly, introducing new and urgent challenges regarding control of apparently new content.
“Most of these problems could be solved relatively quickly, which introduces new and urgent challenges regarding the control of apparently new content.”
Quote in the paperNot a substantial claimClassified as probably not a substantial claim.
InterpretationreasoningScientific text generated by ChatGPT will definitely contain errors or misleading information.
“As a result, scientific text generated by ChatGPT will definitely contain errors or misleading information.”
Quote in the paperSubstantial claimProbably not a substantial claim: may not pass S2 (checkable).
InterpretationreasoningChatGPT is not yet ready to be used in the production of scientific articles.
“One of the main reasons why this AI is not yet ready to be used in the production of scientific articles is its lack of ability to evaluate the veracity and reliability of the information that it processes.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing).
Interpretationown dataThe incorrect references provided by ChatGPT represent a key difference with respect to other computational tools such as typical web browsers, which are focused on providing reliable references.
“This is a huge problem with ChatGPT, and it represents a key difference with respect to other computational tools such as typical web browsers, which are focused on providing reliable references for the required information.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).
InterpretationreasoningLosing trust in the accuracy and integrity of scientific research could have a detrimental impact on the progress of science.
“Losing trust in the accuracy and integrity of scientific research could have a detrimental impact on the progress of science.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based). May combine two claims.
InterpretationreasoningOne of the biggest dangers is the potential for the proliferation of false information in scientific articles, which could lead to a devaluation of the scientific enterprise itself.
“One of the biggest dangers is the potential for the proliferation of false information in scientific articles, which could lead to a devaluation of the scientific enterprise itself.”
Quote in the paperNot a substantial claimClassified as probably not a substantial claim.
InterpretationreasoningChatGPT lacks the ability to evaluate the veracity and reliability of the information it processes.
“One of the main reasons why this AI is not yet ready to be used in the production of scientific articles is its lack of ability to evaluate the veracity and reliability of the information that it processes.”
Quote in the paperSubstantial claim
The paper relies on this without support.
Reviewers may find it not a trivial matter to distinguish between an article written by a human or by ChatGPT.
“It is also important to note that reviewers may find it not a trivial matter to distinguish between an article written by a human or by this AI.”
The paper relies on this without support.
Predatory journals may exploit the opportunity for quick production of scientific articles to generate large amounts of low-quality content.
“A real risk is that predatory journals may exploit the opportunity for the quick production of scientific articles to generate large amounts of low-quality content.”
The paper relies on this without support.
Predatory journals are often motivated by profit rather than a commitment to scientific advancement.
“These journals, often motivated by profit rather than a commitment to scientific advancement, may use AI to rapidly produce articles, flooding the market with subpar research that undermines the credibility of the scientific community.”
The paper relies on this without support.
Predatory journals may use AI to rapidly produce articles, flooding the market with subpar research that undermines the credibility of the scientific community.
“These journals, often motivated by profit rather than a commitment to scientific advancement, may use AI to rapidly produce articles, flooding the market with subpar research that undermines the credibility of the scientific community.”
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
Effective implementation of AI in the pharmaceutical industry relies on access to vast amounts of high-quality patient data.
“Effective implementation of AI in the pharmaceutical industry relies on access to vast amounts of high-quality patient data (Blanco-González et al., 2023).”
From Exploring the antecedents of AI adoption for effective HRM practices in the Indian pharmaceutical sector (2023) · cites this paper as Blanco-González et al., 2023
The citation lists haven’t been fetched yet
References 52 · Cited by 690
Research works printed but not linked: 1
- Schraagen, J.M., and van Diggelen, J. (2021). Expertise at Work, Palgrave Macmillan.
- Authors
- Blanco-González, A., Cabezón, A., Seco-González, A., Conde-Torres, D., Antelo-Riveiro, P., Piñeiro, Á., & Garcia-Fandino, R.
- Affiliations
- Universidade de Santiago de Compostela (ES)
- Department of Organic Chemistry, Center for Research in Biological Chemistry and Molecular Materials, University of Santiago de Compostela, CIQUS, 15705 Santiago de Compostela, Spain
- Corresponding author
- Piñeiro, Ángel
- ORCID
- Venue
- Pharmaceuticals
- Publisher
- MDPI AG
- Published
- 2023-06-18
- Volume
- 16
- Issue
- 6
- Pages
- 891
- ISSN
- 1424-8247 (ISSN-L)
- PMC
- PMC10302890
- OpenAlex
- W4381252028
- Cited by
- 690
- References
- 52
- Citations by year
- 2026: 161 · 2025: 251 · 2024: 207 · 2023: 69 · 2022: 1
- FWCI
- 11.68
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- Pharmaceuticals (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- arXiv (Cornell University) (cc-by)
- Funders
- Spanish Agencia Estatal de Investigación (AEI) and the ERDF (RTI2018-098795-A-I00, PID2019-111327GB-I00, PDC2022-133402-I00, ED431F 2020/05, ED431B 2022/36, 06_IN606D_2021_2600276, ED431G/09, RYC-2016-20335, ED481A 2022/290)
- Xunta de Galicia and the ERDF (RTI2018-098795-A-I00, PID2019-111327GB-I00, PDC2022-133402-I00, ED431F 2020/05, ED431B 2022/36, 06_IN606D_2021_2600276, ED431G/09, RYC-2016-20335, ED481A 2022/290)
- Centro Singular de Investigación de Galicia (RTI2018-098795-A-I00, PID2019-111327GB-I00, PDC2022-133402-I00, ED431F 2020/05, ED431B 2022/36, 06_IN606D_2021_2600276, ED431G/09, RYC-2016-20335, ED481A 2022/290)
- Ministerio de Ciencia, Innovación y Universidades for a “Ramón y Cajal” contract (RTI2018-098795-A-I00, PID2019-111327GB-I00, PDC2022-133402-I00, ED431F 2020/05, ED431B 2022/36, 06_IN606D_2021_2600276, ED431G/09, RYC-2016-20335, ED481A 2022/290)
- Xunta de Galicia for a predoctoral contract (RTI2018-098795-A-I00, PID2019-111327GB-I00, PDC2022-133402-I00, ED431F 2020/05, ED431B 2022/36, 06_IN606D_2021_2600276, ED431G/09, RYC-2016-20335, ED481A 2022/290)
- Topics
- Physical Sciences › Computer Science › Computational Theory and Mathematics › Computational Drug Discovery Methods
- Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
- Physical Sciences › Computer Science › Artificial Intelligence › Artificial Intelligence Applications
- Keywords
- artificial intelligence, drug discovery, data quality, ethical considerations, explainable AI, data augmentation, ChatGPT, AI limitations
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-10.
Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-10.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
