Introduction
The traditional drug discovery process has long been characterized by a significant investment of both time and financial resources. Historically, the journey from the identification of a potential therapeutic target to an approved drug can span more than a decade, costing upwards of billions of dollars. The process typically involves high-throughput screening of large libraries of compounds, followed by preclinical testing and several phases of clinical trials. Despite these efforts, a vast majority of potential drug candidates fail in the later stages, often due to unforeseen toxicity or inefficacy, leading to an immense attrition rate in pharmaceutical research and development (R&D)[1]. I remember talking to one of my mentors, Dr. Guillermo Rodriguez, when he said that out of 10 molecules that enter the pipeline of a pharmaceutical company, only one molecule reaches the commercial market.
In recent years, Artificial Intelligence (AI) has emerged as a powerful tool poised to revolutionize this costly and labor-intensive process. By leveraging machine learning (ML) algorithms and vast datasets, AI has the potential to expedite various stages of drug discovery, from target identification to lead compound optimization. AI can rapidly sift through enormous amounts of chemical and biological data, predicting the properties of molecules and their interactions with biological systems with unprecedented accuracy. This ability not only reduces the time needed for preliminary research but also significantly lowers costs by identifying high-potential drug candidates earlier in the process.
The purpose of this paper is to review some of the multiple topics one can find in the literature surrounding the development of one new drug using Artificial Intelligence (AI) Techniques, the complexity, the main software and companies working in this area, the challenges and the market potential of this new trend in the pharmaceutical industry.
For this article, read over 30 studies, articles, web pages to try to summarize the most important points surrounding the importance of AI use in the new development of medicines.
a. How Does AI Work in Drug Development?
AI-driven drug development relies on advanced algorithms that learn from historical data, such as the molecular properties of compounds, their interactions with biological targets, and the outcomes of previous experiments.
A great variety of systems biology and machine learning approaches are continuously enhanced to accelerate the path to efficient drug development[2].
These algorithms can then predict how new molecules will behave, suggesting potential drug candidates that might have been overlooked by traditional methods.
Moreover, AI can model the chemical and physical properties of drugs in silico, offering predictions about solubility, toxicity, and efficacy without the need for initial lab work.
As AI algorithms become more sophisticated, they are increasingly being integrated into various stages of pharmaceutical R&D, where they assist in everything from chemical design to clinical trial optimization.
According to Kit Kay Mark[3], the AI drug development is described as the use of techniques that enable computers to mimic human behavior (See Figure 1), using machine learning (ML) statistical methods with the ability to learn with or without being explicitly programmed.
Figure 1. AI Drug Development. From Kit-Kat Mark 3.
b. Drugs Developement with AI
The drug discovery process consists of identifying a biological target, designing a molecule capable of interacting with it, optimizing that molecule to enhance its effectiveness, and evaluating its safety and efficacy. Incorporating AI technology into this pipeline can accelerate each of these stages. In the initial phase, AI can analyze vast datasets, such as genetic and protein data, to uncover potential drug targets. Machine learning (ML) algorithms can detect patterns in the data that may go unnoticed by humans, potentially leading to the identification of novel or previously overlooked targets.
Once a target is identified, AI/ML algorithms can assist in designing molecules to interact with it. This involves generating numerous molecular candidates and using these algorithms to select the most promising ones. Researchers can then further refine these candidates by predicting their efficacy, toxicity, pharmacokinetics, and safety in vivo—well before clinical trials begin[4].
Some examples of AI-developed drugs are already showing promise. Exscientia (https://www.exscientia.com), a pioneering AI-driven company, with several options using AI in cases of PRECISION TARGET, PRECISION DESIGN and PRECISION MEDICINE. In this last example, the company´s EXALT-1 Clinical Trial, demonstrated for the first time that the functional precision oncology platform can improve patient outcome in a prospective interventional trial[5].
Another example is Insilico Medicine (https://insilico.com), achieved a milestone in artificial intelligence drug discovery – bringing the first drug discovered and designed by generative AI into Phase II clinical trials with patients. This lead program, for a potentially first-in-class pan-fibrotic inhibitor known as INS018_055 – that demonstrates the validity of Insilico’s end-to-end AI drug discovery platform[6].
Both cases highlight the ability of AI to accelerate drug discovery and deliver promising candidates for clinical evaluation.
c. Clinical Trials of Drugs Currently Studied
Several AI-developed drugs have progressed to clinical trials, marking a significant leap from theoretical models to real-world application.
The first drug designed with the help of AI to enter clinical trials was DSP-1181. This drug was created by British start-up Exscientia and Japanese pharmaceutical firm Sumitomo Dainippon Pharma, and the indication is to treat obsessive compulsive disorders[7].
The compound was discovered by algorithm-based candidate screening and managed to reach clinical testing in 12 months, as opposed to the typical 5 years seen with traditional methods. However, the drug did not progress past Phase I. In July 2022, it was discontinued as it did not achieve the evaluation criteria during its Phase I study[8].
Another medicine was made by Insilico Medicine, a biotech company headquartered in Hong Kong, has created the world’s first AI-designed anti-fibrotic small molecule inhibitor drug to be tested in human patients. Distinguishing itself from other AI-driven medications currently undergoing testing, INS018-055 was both discovered and designed using AI. In June 2023, Phase II, double-blind, randomized clinical trials began. As of August 2023, the trials are currently being conducted in both the U.S. and China, to investigate the safety, tolerability, pharmacokinetics, and efficacy of INS018_055 administered orally in subjects with idiopathic pulmonary fibrosis (IPF) 8.
Over the past decade, following the introduction of these techniques, the number of AI-discovered drug and vaccines molecules has increased substantially. In 2022, we showed that AI-discovered small-molecule numbers were growing exponentially and beginning to match the number of classically discovered small molecules[9].
This trend has continued since then. Similar exponential growth can be seen in AI-discovered biologics, although the number of molecules is still smaller. We also see signs that AI is beginning to accelerate drug discovery timelines9.
These examples underscore the potential of AI not only in accelerating drug discovery but also in expanding the therapeutic landscape by repurposing existing molecules for novel diseases.
d. AI Applications in Drug Discovery
The integration of artificial intelligence (AI) into the pharmaceutical industry has fundamentally changed how new drugs are discovered. AI’s ability to analyze large datasets, identify patterns, and predict molecular behavior has made it an invaluable tool in each stage of the drug discovery process. From identifying biological targets to optimizing lead compounds, AI is rapidly enhancing the efficiency, speed, and precision of drug development.
AI plays a crucial role in target identification, one of the earliest stages of drug discovery. Traditionally, identifying a viable biological target—such as a protein or gene implicated in a disease—was a laborious process, requiring extensive laboratory experiments and data analysis. AI systems, however, can sift through vast genomic, proteomic, and clinical datasets to pinpoint potential drug targets more rapidly[10].
Machine learning algorithms are used to map out the biological pathways and interactions involved in a disease, identifying targets that may not be obvious through traditional methods.
e. The design of new drug molecules
Historically involved trial and error, with chemists and investigators synthesizing compounds and testing them in vitro and in vivo. AI has significantly shortened this process through predictive modeling. AI algorithms can generate molecular structures and predict their biological activity, solubility, toxicity, and overall efficacy before the first compound is synthesized[11]. By using AI, researchers can explore a wider chemical space, leading to the discovery of innovative compounds that might not have been found using traditional techniques.
Once a lead compound has been identified, it must be refined to improve its efficacy, safety, and pharmacokinetic properties. AI excels in this optimization phase by analyzing the structural activity relationship (SAR) of compounds and predicting modifications that will enhance their performance[12]. Machine learning models trained on historical data can suggest modifications that will increase a compound’s potency, reduce its toxicity, or improve its ability to penetrate biological barriers like the blood-brain barrier.
One notable example of AI’s success in lead optimization is Insilico Medicine (www.insilico.com), which utilized AI algorithms to discover and optimize a lead compound for idiopathic pulmonary fibrosis (IPF) in just 18 months—significantly faster than traditional drug development timelines.
AI is revolutionizing the prediction of a drug’s ADMET® properties—Absorption, Distribution, Metabolism, Excretion, and Toxicity—which are critical for determining whether a drug candidate is likely to be safe and effective in humans. AI models can be trained on large datasets of existing drugs and their ADMET profiles to predict how new compounds will behave in vivo[13]. These predictions allow pharmaceutical companies to rule out problematic compounds early in the process, reducing the risk of failure in later stages of development.
For instance, AI models developed by Atomwise (www.atomwise.com) and Schrödinger (www.schrodinger.com) have been successfully used to predict drug metabolism and toxicity profiles, enabling the rapid exclusion of compounds with unfavorable characteristics. This early filtering of candidates reduces costly failures in clinical trials, thus lowering overall R&D expenditure.
Also, there are other ways to identify bioactive small molecules for clinical studies, ones is the High-throughput screening (HTS). HTS, as the name indicates, is a drug discovery process that enables a biochemical or cellular event to be reproducibly and rapidly tested against chemical entities many hundreds of thousands of times[14].
By integrating computational methods with expansive on-demand chemical libraries, it is now possible to access a significantly broader chemical space, contingent on the predictive accuracy of these approaches.
f. AI in Preclinical and Clinical Trials
As mentioned before, AI is also transforming preclinical and clinical trials, two critical stages in drug development where success or failure can determine whether a drug reaches the market. Traditionally, these phases have been costly, time-consuming, and fraught with a high risk of failure. AI is now being employed to modernize these processes, improving trial design, patient selection, and real-time data analysis, all of which reduce risks, save time, and cut costs.
Preclinical trials currently involve the testing of a drug on cell cultures and animal models to determine safety, dosage, and pharmacological activity (Clinical studies Phase I and II). This stage requires extensive data collection and analysis, which AI can enhance by rapidly processing large volumes of experimental data. Machine learning models can predict how a compound will behave in vivo based on preclinical data, providing researchers with early insights into potential issues such as toxicity or poor absorption[15].
AI-powered platforms, such as those developed by DeepMind® and Schrödinger®, are increasingly being used to model complex biological processes, including protein folding and molecular dynamics. These models help predict how drug candidates will interact with biological targets, thereby refining compound selection for further development. By using AI to process and interpret vast datasets from preclinical trials, pharmaceutical companies can focus their efforts on compounds with the greatest likelihood of success.
AI has demonstrated great potential in revolutionizing the design and management of clinical trials. One of the key advantages of AI is its ability to optimize clinical trial designs by simulating different trial parameters, such as dosing regimens, patient populations, and endpoints. AI algorithms can analyze historical clinical trial data to suggest optimal trial designs that maximize the chances of success while minimizing risks. For example, AI can aid in the selection of biomarkers that predict patient responses to treatment, which can lead to more targeted and personalized clinical trials[16].
AI-driven platforms like IQVIA’s Virtual Trials® (https://www.iqvia.com/solutions/research-and-development) and Phesi® (https://www.phesi.com) are being used to improve trial designs, reduce the likelihood of trial failure, and identify patient subgroups most likely to benefit from new therapies.
One of the greatest challenges in clinical trials is identifying and recruiting suitable patients. AI can address this issue by analyzing electronic health records (EHRs), genomic data, and other real-world data sources to find patients who meet the inclusion and exclusion criteria for a specific trial. AI models can sift through enormous datasets to match patients based on genetic markers, disease progression, and other clinical factors that might not be immediately apparent through traditional screening methods.
By improving patient selection, AI not only accelerates recruitment but also helps ensure that clinical trials are more personalized and effective. For example, IBM Watson Health has developed AI-based tools that help identify eligible patients for oncology trials, reducing recruitment times and ensuring that trials proceed more efficiently.
g. Once clinical trials are underway
AI enables real-time data monitoring, allowing researchers to make informed decisions based on patient responses and potential adverse events. This real-time analysis can result in adaptive trials, where protocols are adjusted based on intermediate results[17]. AI can detect patterns in the data early, identifying safety signals or efficacy trends before they would become apparent through traditional statistical methods.
For instance, AI-driven platforms like Medidata® from Dassault Systems, creators of Solidworks® (www.medidata.com) use machine learning to analyze real-time data from ongoing trials, providing insights into patient safety and treatment efficacy. This capability allows for more adaptive trial designs, where patient cohorts or dosing regimens can be modified based on initial findings, thereby reducing trial length and increasing the likelihood of success.
h. The regulatory landscape
Surrounding AI-driven drug discovery is still evolving, and there is a lack of clear guidelines from regulatory bodies such as the U.S. Food and Drug Administration (FDA) or the European Medicines Agency (EMA). Traditionally, regulatory approval processes are based on well-established experimental protocols, and introducing AI raises questions about validation and accountability.
One of the key concerns is how to assess the reliability of AI-generated drug candidates. Regulators need to be confident that the AI models are robust and that their predictions are reproducible.
Ensuring the reliability and robustness of AI solutions, whether for internal stakeholders or external clients, requires a structured and transparent approach.
Regulatory authorities in the EU, US, and UK are increasingly addressing the role of AI and ML in drug development. In the EU, the European Medicines Agency (EMA) has taken a proactive step by issuing a draft reflection paper that outlines how AI and ML should be integrated across the drug lifecycle. This paper promotes a human-centric and risk-based approach, holding marketing authorization (MA) applicants and holders responsible for ensuring compliance with existing regulations, such as GxP standards[18]. The EMA emphasizes the role of AI in enhancing drug development stages from discovery through post-authorization, with varying risk levels depending on AI’s use, such as lower risks in drug discovery and higher risks in precision medicine[19]
The US FDA has also explored AI’s potential through its 2023 discussion paper, which aims to foster dialogue on the benefits and challenges AI presents in drug development. The FDA highlights AI’s potential in drug discovery, clinical trials, and predictive modeling but raises concerns about the introduction of risks such as data biases and lack of transparency in AI models[20]. To address these challenges, the FDA emphasizes the need for standards focused on reliability, safety, and bias mitigation, ensuring that AI’s use does not compromise the integrity of the drug development process[21]
In the UK, the Medicines and Healthcare Products Regulatory Agency (MHRA) has not yet published formal guidance on AI in the drug lifecycle, but its future regulations are expected to align with the UK’s AI White Paper. This would reflect the UK’s sector-specific, principles-based approach to AI regulation. Additionally, the MHRA’s ongoing collaboration with international regulatory bodies such as the US FDA and Health Canada on machine learning practices for medical devices will likely influence future guidelines for AI in drug development19.
i. Challenges and limitations of ai in drug discovery
We have seen through this point in the article that while artificial intelligence (AI) is transforming the drug discovery process, it is not without its challenges and limitations. These obstacles must be addressed to unlock the full potential of AI in pharmaceutical research and development. The major challenges include issues related to data quality, regulatory hurdles, model interpretability, and integration into existing workflows.
AI systems thrive on large datasets, but the quality and availability of relevant data can often be problematic in drug discovery. For AI algorithms to produce meaningful results, they require vast amounts of high-quality, well-annotated, and standardized data. However, much of the data in pharmaceutical research is fragmented, incomplete, or proprietary, making it difficult for AI to deliver accurate predictions.
Biological data can be highly variable across different research institutions and experimental conditions. For example, genomic data might differ due to variations in sample collection, sequencing techniques, or data preprocessing methods. This inconsistency can limit the performance of AI models, as they struggle to generalize across disparate datasets. “An immense challenge—one of the most central facing 21st century biology—is that of managing the variety and complexity of data types, the hierarchy of biology, and the inevitable need to acquire data by a wide variety of modalities”[22]. Ensuring that data is reliable and comparable is essential for AI to become a mainstay in drug discovery.
It is critical to establish and agree upon a measurable accuracy threshold with the client. AI models must undergo extensive testing—pushing the limits to uncover potential weaknesses. This testing process should involve the collection and integration of a comprehensive range of target data during model training, with particular attention given to identifying and managing flaws early in the process[23].
AI systems should be designed with the capability to recognize and adapt to target content that was not included in the original training data. This adaptability ensures the system’s nondeterministic nature and its potential to generalize beyond the initial dataset[24].
Finally, exceeding client expectations in accuracy is key. A prudent solution provider should aim for internal testing outcomes that surpass the client’s required accuracy by at least 4%. This additional margin accounts for unexpected variability in client-specific data, ensuring that the AI solution remains robust and reliable in real-world applications[25].
This requires the development of new regulatory frameworks specifically tailored to AI-based drug discovery. Pharmaceutical companies must work closely with regulatory agencies to establish these guidelines and ensure the safe and ethical application of AI.
Another challenge is the issue of model interpretability. Many AI models, particularly those based on deep learning, operate as “black boxes,” meaning that while they can make accurate predictions, their inner workings are not easily understood by researchers.
In AI, a “black box” refers to a model or system whose internal workings are not easily understood or accessible to users[26]. While the AI can make accurate predictions or decisions, it does so in a way that is difficult to interpret or explain. Users can see the input and the output, but the decision-making process remains opaque. This lack of transparency makes it challenging to understand how or why the AI arrived at a particular conclusion, raising concerns about trust, accountability, and the ability to verify its decisions.
This lack of transparency makes it difficult to explain how or why certain predictions are made, which can lead to hesitancy in adopting AI in highly regulated environments such as pharmaceutical research[27].
The increasing deployment of opaque AP applications in healthcare, has amplified on the capacity of this models to be both explainable and interpretable to gain the trust of healthcare professionals to trust the decision-making process and under lying logic behind the software.
Pharmaceutical companies face challenges when attempting to integrate AI into their traditional research and development (R&D) workflows. AI-driven drug discovery requires not only the adoption of new technologies but also a cultural shift within organizations[28]. Many researchers are trained in traditional methods and may be reluctant to embrace AI, especially if they do not fully understand how the technology works or are concerned about job displacement.
Integrating AI tools into existing processes can be complex, as it requires harmonizing AI platforms with legacy systems for data management, laboratory information, and clinical trial design[29]. The successful integration of AI into drug discovery will depend on how seamlessly it can be combined with established workflows while complementing, rather than replacing, the expertise of scientists.
The use of AI in drug discovery also raises important ethical considerations. One major concern is bias in AI algorithms, which could lead to unequal drug development outcomes[30]. For example, if AI models are trained on datasets that are not representative of diverse populations, there is a risk that the resulting drugs may not be effective or safe for underrepresented groups. Addressing this issue requires conscious efforts to ensure diversity and inclusivity in the data used to train AI systems.
Additionally, transparency is critical when using AI in drug discovery. Pharmaceutical companies must ensure that AI algorithms are applied ethically and that there is accountability for the decisions made by these models. Establishing clear guidelines for the ethical use of AI, including data privacy, algorithmic bias, and patient safety, is essential to building trust in AI-driven drug development[31].
j. The Market
According to marketsandmarkets.com, the AIdrug discovery market mainly driven by players like NVIDIA, EXSCIENTIA and GOOGLE, will grow from sales of $0.9billion USD in 2023, to $4.9 billion in 2028 with a Compound Annual Growth Rate of 40.2%.
This is a fraction of the total healthcare AI industry that is valued in $20.9 billion USD in 2024. The drug discovery market is mainly driven by companies like NVIDIA, Exscientia and Google.[32].
According to BIOSPACE, one newspaper dedicated to topics in health and science, the global AI in drug discovery market was evaluated at US$ 1.70 billion in 2023 and is expected to attain around US$ 11.93 billion by 2033, growing at a CAGR of 21.5% from 2024 to 2033[33].
Despite the differences between estimations, the fact is that this is a strong industry that is accelerating and transforming workflows, processes that in the bottom line, reduce the process decision making in the drug development industry.
This market is experiencing a predictable increase in mergers and acquisitions (M&A) activity. Major pharmaceutical companies and technology giants are actively acquiring AI startups and smaller companies with specialized expertise.
According to a report from Grand View Research, BioNTech acquired InstaDeep, a global AI and machine learning technology company. This consolidation is fostering collaboration, accelerating innovation, and leading to the development of more effective and efficient drug discovery solutions[34].
Karen Madden, Chief Technology Officer for the Life Science business sector of Merck, indicate that: «With millions of people waiting for the approval of new medicines, bringing a drug to market, still takes on average, more than 10 years and costs over 1.9 Million Euros, Our platform enables any laboratory to count on generative AI to identify the most suitable drug-like candidates in a vast chemical space. This helps ensure the optimal chemical synthesis route for development of a target molecule in the most sustainable way possible.26”
k. Conclusion
The integration of artificial intelligence (AI) into drug discovery is revolutionizing the pharmaceutical industry by improving efficiency across the drug development lifecycle, from preclinical studies to clinical trials. The largest share of the AI-driven drug discovery market is held by preclinical and clinical testing, where AI technologies are now indispensable. AI enhances drug discovery through better understanding of molecular mechanisms, optimizing dose-response predictions, and improving clinical trial success rates. Applications of AI span various therapeutic areas, with immuno-oncology and neurodegenerative disorders leading the way.
AI’s ability to handle vast datasets, analyze complex genomic and proteomic information, and identify novel drug candidates has proven especially valuable in high-demand areas such as cancer treatment and neurodegenerative diseases like Alzheimer’s. Biotechnology and pharmaceutical companies have recognized the transformative potential of AI, heavily investing in these technologies to accelerate drug discovery and reduce costs.
However, data-related challenges, such as handling diverse and uncertain pharmaceutical datasets, remain a significant obstacle. Addressing these challenges is crucial to realizing AI’s full potential in drug discovery. As AI continues to evolve, its application in drug formulation and manufacturing also presents new opportunities for efficiency and innovation, further driving growth in the pharmaceutical industry.
[1] Guillermo Rodríguez. Manual de Investigación Clínica. Segunda Edición. 979-8579174222
[2] Clemence Reda. Machine learning applications in drug development. Computational and Structural Biotechnology Journal. Volume 18. 2020
[3] Kit-KAI Mark. Artificial intelligence in drug development: present status and future prospect. Drug Discovery Today. Volume 24, Issue 3. March 2019
[4] Anita Ioana Visan and Irina Negut. Integrating Artificial Intelligence for Drug Discovery in the Context of Revolutionizing Drug Delivery. 2024 Feb; 14(2): 233. Published online 2024 Feb 7
[5] Kornauth C, Pemovska T. Et Al. Functional Precision Medicine Provides Clinical Benefit in Advanced Aggressive Hematologic Cancers and Identifies Exceptional Responders. Cancer Discov. 2022 Feb;12(2):372-387.
[6] https://insilico.com/blog/first_phase2
[8] https://vial.com/blog/articles/what-ai-designed-drugs-have-been-fda-approved/?
[9] Madura KP Jayatunga. How successful are AI-discovered drugs in clinical trials? A first analysis and emerging lessons. Drug Discovery Today. Volume 29, Issue 6, June 2024, 104009
[10] Shruti Singh Et Al. Artificial Intelligence and Machine Learning in Pharmacological Research: Bridging the Gap Between Data and Drug Discovery. Cureus. 2023 Aug 30;15(8):e44359.
[11] Paul D, Sanap G Et Al. Artificial intelligence in drug discovery and development. Drug Discov Today. 2021 Jan;26(1):80-93. doi: 10.1016/j.drudis.2020.10.010. Epub 2020 Oct 21.
[12] Vora LK, Gholap Et Al.. Artificial Intelligence in Pharmaceutical Technology and Drug Delivery Design. Pharmaceutics. 2023 Jul 10;15(7):1916
[13] Kyle Swanson. ADMET-AI: a machine learning ADMET platform for evaluation of large-scale chemical libraries. Bioinformatics, Volume 40, Issue 7, July 2024, btae416
[14]The Atomwise AIMS REPORT. “AI is a viable alternative to high throughput screening: a 318-target study”. Scientific Reports volume 14, Article number: 7526 (2024)
[15] Siyun Yang, Supratik Kar. “Application of artificial intelligence and machine learning in early detection of adverse drug reactions (ADRs) and drug-induced toxicity”. Artificial Intelligence Chemistry. Volume 1, Issue 2. , December 2023.
[16] Shuroug A. Alowais ET AL. “Revolutionizing healthcare: the role of artificial intelligence in clinical practice”. BMC Med Educ. 2023; 23: 689
[17] Singh S. Artificial Intelligence and Machine Learning in Pharmacological Research: Bridging the Gap Between Data and Drug Discovery. Cureus. 2023 Aug 30;15(8):e44359.
[18] EMEA. “Reflection paper on the use of artificial intelligence in the lifecycle of medicines”. 19 July 2023
[19] “Accelerating how new drugs are made with machine learning. ScienceDaily”. University of Cambridge. The original text of this story is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
[20] FDA. “Artificial Intelligence and Machine Learning (AI/ML) for Drug Development”. 03/18/2024
[21] Ekaterina Pesheva. “Can AI transform the way we discover new drugs?”. Harvard Medical School. November 17, 2022
[22] John C Wooley and Herbert S Lin. “Catalyzing Inquiry at the Interface of Computing and Biology”. Washington (DC): National Academies Press (US); 2005.
[23] https://www.oracle.com/fr/artificial-intelligence/ai-model-training/
[24] Lalli Myllyaho. Systematic literature review of validation methods for AI systems. Journal of Systems and Software. Volume 181. , November 2021, 111050
[25] Jon Patrick . “How to Check the Reliability of Artificial Intelligence Solutions—Ensuring Client Expectations are Met”. Appl Clin Inform. 2019 Mar; 10(2): 269–271. 2019.
[26] “What is black box AI?” A guide to artificial intelligence in the enterprise
[27] Shin D. “User perceptions of algorithmic decisions in the personalized ai system: perceptual evaluation of fairness, accountability, transparency, and explainability”. J Broadcast Electron Media. 2020;64(4):541–65.
[28] Brian Buntz. Pharma’s cultural barriers could blunt AI progress despite pressures for more efficient drug approvals. Drug Discovery and development. May 8 2024.
[29] Navigating the Complexity: Streamlining Integration Of AI Solutions With Existing Systems. March 16, 2024in AI, Digital Enterprise, Infrastructure, Research & Development
[30] Mitul Harishbhai Tilala Et al. Ethical Considerations in the Use of Artificial Intelligence and Machine Learning in Health Care: A Comprehensive Review. 2024 Jun; 16(6)
[31] Madhan Jeyaraman Et Al. Unraveling the Ethical Enigma: Artificial Intelligence in Healthcare
2023 Aug; 15(8): e43262.
[32] http://www.marketsandmarkets.com
[33] https://www.biospace.com/ai-in-drug-discovery-market-size-to-expand-us-11-93-bn-by-2033#:~:text=The%20global%20artificial%20intelligence%20(AI,21.5%25%20from%202024%20to%202033.
[34] Artificial Intelligence In Drug Discovery Market Size, Share & Trends Analysis Report By Therapeutic Space (Oncology, Neurodegenerative Diseases), By Application, By Region, And Segment Forecasts, 2024 – 2030. Grand view research
aiglobalhealth.blog 
Deja un comentario