Can we open the black box of AI?
Open accessThe DOI matches its Crossref record; 5 of the 6 references it cites were found in a registry or an index.
Abstract
Artificial intelligence is everywhere. But before scientists trust it, they first need to understand how machines learn.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
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?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“most contemporary AI models operate as black boxes, making their internal processes difficult to comprehend (Castelvecchi 2016).”
From Artificial intelligence for literature reviews: opportunities and challenges (2024) · cites this paper as Castelvecchi 2016
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“The latter space comprises hundreds of layers and millions of parameters, which makes DNNs be considered as complex black-box models [4].”
From Explainable Artificial Intelligence (XAI): Concepts, taxonomies, opportunities and challenges toward responsible AI (2020) · cites this paper as [4]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“Lastly, a prominent concern of AI models based on neural network architecture is their often criticized lack of interpretability which earned them the term ‘black boxes’27,28.”
From Large language models and multimodal foundation models for precision oncology (2024) · cites this paper as [27]
The cited source says what the claim attributes to it.
In the citing paper
“Most of these discussions point to the fact that algorithmic decision-making is often difficult to understand for - and explain to - human agents, due to the different, and sometimes inscrutable for persons, ways of AI operations: the so-called 'black box' problem (Castelvecchi, 2016).”
From Four Responsibility Gaps with Artificial Intelligence: Why they Matter and How to Address them (2021) · cites this paper as Castelvecchi, 2016
The cited source says what the claim attributes to it.
In the citing paper
“Not being able to understand what has been captured by the model [4] and stored in its internal representation may entail a privacy breach.”
From Explainable Artificial Intelligence (XAI): Concepts, taxonomies, opportunities and challenges toward responsible AI (2020) · cites this paper as [4]
2 more claims about this paper
Create a free account to see every claim other papers make about this one.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
ImageNet classification with deep convolutional neural networks
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2017) · Communications of the ACM · doi:10.1145/3065386 · Cited by 112,155
Advances in neural information processing systems 7
Cios, K., & Shields, M. (1997) · Neurocomputing · doi:10.1016/s0925-2312(97)90018-7 · Cited by 22,059
Advances in Neural Information Processing Systems 25
Zhang, Y., Sutton, C. A., Storkey, A., & Zoubin Ghahramani, A. (2012) · Cited by 12,653
Distilling Free-Form Natural Laws from Experimental Data
Schmidt, M., & Lipson, H. (2009) · Science · doi:10.1126/science.1165893 · Cited by 2,890
Probabilistic machine learning and artificial intelligence
Ghahramani, Z. (2015) · Nature · doi:10.1038/nature14541 · Cited by 1,995
Cited by
Items 901–1,000 of 1,006
Interpretable deep learning for deconvolutional analysis of neural signals
Tolooshams, B., Matias, S., Wu, H., Temereanca, S., Uchida, N., Murthy, V. N., Masset, P., & Ba, D. E. (2024) · Neuron · doi:10.1016/j.neuron.2025.02.006 · Cited by 12
“42,43 Our method complements these tools by extracting interpretable impulse-like responses from multiplexed signals in single neurons, recovering locally low-rank structures to characterize heterogeneity and homogeneity across neural populations.”
Du, R. F., Carbonell, E. L., Huang, J., Liu, S., Wang, X., Shen, D. G., & Ke, J. (2025) · IEEE Transactions on Medical Imaging · doi:10.1109/tmi.2025.3551913 · Cited by 19
OmniXAS: A universal deep-learning framework for materials x-ray absorption spectra
Kharel, S. R., Meng, F., Qu, X., Carbone, M. R., & Lu, D. (2025) · Physical Review Materials · doi:10.1103/physrevmaterials.9.043803 · Cited by 19
Cyber-governance of the natural world the implications of digital twins in environmental governance
Korenhof, P., Metcalfe, B., & Henschke, A. (2025) · Earth System Governance · doi:10.1016/j.esg.2025.100276 · Cited by 21
Tugarin, N., & van Husen, C. (2025) · International Journal of Information Management Data Insights · doi:10.1016/j.jjimei.2025.100357 · Cited by 19
AI, originality, and attribution: Researchers’ perspectives on distinguishing contributions
Wu, Y., Lu, X., & Lin, C. (2025) · Accountability in Research · doi:10.1080/08989621.2025.2536817 · Cited by 9
“Concurrently, the deployment of highly complex, often opaque “black box” algorithms complicates the scientific imperative for transparency and reproducibility (Castelvecchi 2016; Rudin 2019).”
Wu, X., Wang, X., Kim, T., Saleh, K., Fu, H., & Xiong, C. (2025) · Transportation Research Part E Logistics and Transportation Review · doi:10.1016/j.tre.2025.104283 · Cited by 8
Söderlund, K. (2025) · Technology and Regulation · doi:10.71265/6bedar76 · Cited by 20
Riemer, K., Peter, S., Schwabe, G., Chatterjee, S., Adam, M., & Davison, R. M. (2025) · Information Systems Journal · doi:10.1111/isj.70027 · Cited by 13
“This also limits scrutiny, as it is typically infeasible to attribute a response to specific training items (Castelvecchi 2016).”
AI-Driven Chemometrics for Multi-omics Data Integration: Advances, Challenges, and Future Directions
Polu, P. R. (2026) · Critical Reviews in Analytical Chemistry · doi:10.1080/10408347.2026.2657553 · Cited by 14
Perez-Cerrolaza, J., Flores, J. L., Blum, C., Cerquides, J., & Abuin, A. (2021) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2021.3074394 · Cited by 8
The differential diagnosis of IgG4-related disease based on machine learning
Yamamoto, M., Nojima, M., Kamekura, R., Souta-Kuribara, A., Uehara, M., Yamazaki, H., Yoshikawa, N., Aochi, S., Mizushima, I., Watanabe, T., et al. (2022) · Arthritis Research & Therapy · doi:10.1186/s13075-022-02752-7 · Cited by 9
“The cutoff values for the age at the first visit and serum IgA concentration in the Gini impurity are unknown due to the nature of the random forest method [25], but IgG4-RD often occurs in the elderly [26], and age is presumably used to differentiate IgG4-RD from MCD, as described above [24].”
MethodProving properties of binary classification neural networks via Łukasiewicz logic
Preto, S. M. D. S., & Finger, M. (2022) · Logic Journal of IGPL · doi:10.1093/jigpal/jzac050 · Cited by 8
Multi-Information Model for Large-Flowered Chrysanthemum Cultivar Recognition and Classification
Wang, J., Tian, Y., Zhang, R., Liu, Z., Tian, Y., & Dai, S. (2022) · Frontiers in Plant Science · doi:10.3389/fpls.2022.806711 · Cited by 8
“However, because of the difficulties in data acquisition, many deep learning methods used pre-training models based on ImageNet (He et al., 2015, 2016; Russakovsky et al., 2015; Chattopadhay et al., 2018) as the backbone network, but ImageNet-trained CNNs are strongly biased toward recognizing…”
BackgroundMethodDeep Dynamic Patient Similarity Analysis: Model Development and Validation in ICU
Sun, Z., Lü, X., Duan, H., & Li, H. (2022) · Computer Methods and Programs in Biomedicine · doi:10.1016/j.cmpb.2022.107033 · Cited by 11
Ishii-Rousseau, J. E., Nawa, N., Hashimoto, S., Shigemitsu, H., & Fujiwara, T. (2022) · Anaesthesia Critical Care & Pain Medicine · doi:10.1016/j.accpm.2022.101167 · Cited by 10
Fully Connected Multi-Kernel Convolutional Neural Network Based on Alzheimer’s Disease Diagnosis
Deng, L., Initiative, A. D. N., & Wang, Y. (2023) · Journal of Alzheimer s Disease · doi:10.3233/jad-220519 · Cited by 9
“Additionally, in the biomedical field, more and more people consider deep learning models to be “black box” algorithms [19].”
Biedma-Rdguez, C., Gacto, M. J., Anguita‐Ruiz, A., Alcalá, R., Aguilera, C. M., & Alcalá‐Fdez, J. (2023) · Fuzzy Sets and Systems · doi:10.1016/j.fss.2023.03.014 · Cited by 9
Wu, Z., Yao, L., Liu, W., Zhang, S., Zhang, L., Lu, Z., Wang, J., Chen, B., Luo, R., Li, X., et al. (2023) · American Journal of Clinical Pathology · doi:10.1093/ajcp/aqad058 · Cited by 8
Ardito, V., Cappellaro, G., Compagni, A., Petracca, F., & Preti, L. M. (2023) · JMIR Research Protocols · doi:10.2196/47971 · Cited by 10
“Research on ML implementation has been predominantly conceptual in nature, with the identification of distinguishing issues in the development and dissemination in clinical practice, such as the opacity or black-box effect, and concerns surrounding equity, data privacy, and security [7,18].”
Enhancing soft computing techniques to actively address imbalanced regression problems
Arteaga, M., Gacto, M. J., Galende, M., Alcalá‐Fdez, J., & Alcalá, R. (2023) · Expert Systems with Applications · doi:10.1016/j.eswa.2023.121011 · Cited by 11
Enhancing corporate governance: A conceptual approach to artificial intelligence usage
Correia, A. C., & Água, P. B. (2023) · New Outlooks for the Scholarly Research in Corporate Governance · doi:10.22495/nosrcgp23 · Cited by 11
“These issues include privacy, fairness, transparency, and accountability (Castelvecchi, 2016).”
BackgroundChakraborty, C., Bhattacharya, M., Islam, M. A., & Agoramoorthy, G. (2023) · International Journal of Surgery · doi:10.1097/js9.0000000000000701 · Cited by 9
Ethical Challenges of Integrating AI into Healthcare
Lehmann, L. S. (2022) · Springer eBooks · doi:10.1007/978-3-030-64573-1_337 · Cited by 7
Current and future applications of artificial intelligence in multiple sclerosis
Cacciaguerra, L., Storelli, L., Rocca, M. A., & Filippi, M. (2022) · Elsevier eBooks · doi:10.1016/b978-0-323-90037-9.00012-6 · Cited by 10
Biedma-Rdguez, C., Gacto, M. J., Anguita‐Ruiz, A., Alcalá‐Fdez, J., & Alcalá, R. (2022) · International Journal of Fuzzy Systems · doi:10.1007/s40815-022-01324-w · Cited by 10
“This is known as the ‘‘black box problem,’’ which has been recently discussed in the prestigious journal Nature [1].”
BackgroundInfluentialXu, H., Zheng, Q., Zhu, J., Xie, Z., Cheng, H., Li, P., & Ji, Y. (2022) · Disease Markers · doi:10.1155/2022/7593750 · Cited by 9
“In order to trade o ff the performance and interpretability of the models, a large number of researchers have researched on interpretable disease diagnosis models [14, 15, 36], focusing on interpreting deep black-box models [37].”
BackgroundArgumentation-Based Query Answering under Uncertainty with Application to Cybersecurity
Leiva, M., García, A. J., Shakarian, P., & Simari, G. I. (2022) · Big Data and Cognitive Computing · doi:10.3390/bdcc6030091 · Cited by 9
Les quatre éthiques de l’intelligence artificielle
Ménissier, T. (2023) · Revue d anthropologie des connaissances · doi:10.4000/rac.29961 · Cited by 18
Recommender Systems: Legal and Ethical Issues
Genovesi, S., Kaesling, K., & Robbins, S. (2023) · The International library of ethics, law and technology · doi:10.1007/978-3-031-34804-4 · Cited by 15
Gallardo‐Pizarro, A., Peyrony, O., Chumbita, M., Monzó, P., Aiello, T. F., Teijón‐Lumbreras, C., Gras, E., Mensa, J., Soriano, À., & García‐Vidal, C. (2024) · Expert Review of Anti-infective Therapy · doi:10.1080/14787210.2024.2322445 · Cited by 14
“A prevailing sentiment – challenging to alleviate, though – is the unease associated with replacing clinical decision-making processes with opaque decision tools, commonly referred to as ‘black boxes,’ and possible consequences of computational errors [13,16,27–29].”
BackgroundColor-Patterns to Architecture Conversion through Conditional Generative Adversarial Networks
Navarro-Mateu, D., Carrasco, O. L., & Nieves, P. C. (2021) · Biomimetics · doi:10.3390/biomimetics6010016 · Cited by 8
A study on deep learning approaches over Malware detection
Kavitha, P. M., & Muruganantham, B. (2020) · IEEE International Conference on Advances and Developments in Electrical and Electronics Engineering (ICADEE) · doi:10.1109/icadee51157.2020.9368924 · Cited by 7
Deep Learning for Lung Cancer Nodal Staging and Real-World Clinical Practice
Park, C. M., & Lee, J. H. (2021) · Radiology · doi:10.1148/radiol.2021211981 · Cited by 8
Natural Language Explanations for Machine Learning Classification Decisions
Burton, J., Moubayed, N. A., & Enshaei, A. (2023) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn54540.2023.10191637 · Cited by 9
“However, for complicated architectures such as deep neural networks, it is challenging to trace which features were relied upon most for making the decision [6], [7] as these black-box models employ billions of parameters to make predictions, making them difficult to troubleshoot and trust.”
BackgroundModelling the ultimate pullout resistance of geogrids
Marx, D. H., & Zornberg, J. G. (2024) · Geosynthetics International · doi:10.1680/jgein.23.00172 · Cited by 7
Ludeña, A. F. (2024) · European Public & Social Innovation Review · doi:10.31637/epsir-2024-492 · Cited by 15
Statistical analysis of Cu content effects on structural properties in CuZr metallic glasses
Amigó, N. (2024) · Physica Scripta · doi:10.1088/1402-4896/ad79c0 · Cited by 7
Rogstadkjernet, M., Zha, S. Z., Klæboe, L. G., Larsen, C. K., Aalen, J. M., Scheirlynck, E., Singstad, B.-J., Droogmans, S., Cosyns, B., Smiseth, O. A., et al. (2024) · BMC Medical Imaging · doi:10.1186/s12880-024-01470-7 · Cited by 7
“While a few mistakes may not have a large impact on model statistics, this “black box” problem [18] could be significant in a clinical setting if the operator does not understand what is happening or is unable to intervene.”
BackgroundMakanadar, A., & Shahane, S. (2024) · Lecture notes in civil engineering · doi:10.1007/978-981-97-8116-4_2 · Cited by 8
Retrieval Augmented Medical Diagnosis System
Johnson, E. T., Bande, J. K., & Thomas, J. (2025) · Biology Methods and Protocols · doi:10.1093/biomethods/bpaf017 · Cited by 7
Artificial intelligence in early screening for esophageal squamous cell carcinoma
Yan, S.-Y., Fu, X.-Y., Yang, Y., Jia, L.-Y., Liang, J.-W., Li, Y., Yan, L., Zhou, Y., Zhou, X.-B., Li, S.-W., et al. (2025) · Best Practice & Research Clinical Gastroenterology · doi:10.1016/j.bpg.2025.102004 · Cited by 8
Garí‐Galíndez, J., Fonoll‐Rubio, R., Andrade‐Arvizu, J., Witte, W., Pérez-Rodríguez, A., Paetel, S., Vidal‐Fuentes, P., Guc, M., & Izquierdo‐Roca, V. (2025) · Advanced Energy Materials · doi:10.1002/aenm.202502420 · Cited by 9
Enhancing healthcare resilience through AI-driven performance management systems
Cavadi, G., & Cosenz, F. (2025) · Journal of Health Organization and Management · doi:10.1108/jhom-05-2025-0261 · Cited by 12
Baral, A., Chowdhury, M. A., Rafi, F. A., Khosru, M. A. A., Hossain, S., Chowdhury, S., & Islam, M. A. (2025) · Composites Part B Engineering · doi:10.1016/j.compositesb.2025.113310 · Cited by 9
Training Adversarial Agents to Exploit Weaknesses in Deep Control Policies
Kuutti, S., Fallah, S., & Bowden, R. (2020) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra40945.2020.9197351 · Cited by 65
Reading bots: The implication of deep learning on guided reading
Huang, B., Dou, J., & Zhao, H. Y. (2023) · Frontiers in Psychology · doi:10.3389/fpsyg.2023.980523 · Cited by 18
“Deep learning models comprise hundreds of layers and millions of parameters, which makes deep neural networks (DNNs) considered complex black-box models (Castelvecchi, 2016).”
BackgroundAnwer, D., Gubinelli, F., Kurt, Y. A., Sarauskyte, L., Jacobs, F., Venuti, C., Sandoval, I. M., Yang, Y., Stancati, J. A., Mazzocchi, M., et al. (2023) · PLoS ONE · doi:10.1371/journal.pone.0284480 · Cited by 7
“With current knowledge and state of technology, we are unable to provide an explanation for this pattern [84].”
BackgroundWeakly privileged learning with knowledge extraction
Fu, S., Dong, T., Wang, Z., & Tian, Y. (2024) · Pattern Recognition · doi:10.1016/j.patcog.2024.110517 · Cited by 8
Park, Y. W., Eom, S., Kim, S., Lim, S., Park, J. E., Kim, H. S., You, S. C., Ahn, S. S., & Lee, S. (2024) · Computer Methods and Programs in Biomedicine · doi:10.1016/j.cmpb.2024.108288 · Cited by 12
A convolutional neural network model for accurate short-term leaf area index prediction
Jiang, C., Yeddula, S. D., & Ku, W. (2024) · Modeling Earth Systems and Environment · doi:10.1007/s40808-024-02090-4 · Cited by 11
Integrating AI in Psychotherapy: An Investigation of Trust in Voicebot Therapists
Mühl, L., Stecker, L., Herter, E., Szczuka, J. M., Wischnewski, M., & Krämer, N. C. (2024) · Nordic Conference on Human-Computer Interaction · doi:10.1145/3679318.3685353 · Cited by 10
“As the complexity of AI systems and algorithms grows, people increasingly regard them as ”black boxes” that challenge comprehension in the sense that a large amount of skill and information are necessary to comprehend the AI decision or performance [41].”
Weighted Hermite Variable Projection Networks for Classifying Visually Evoked Potentials
Dózsa, T., Böck, C., Meier, J. M., & Kovács, P. (2024) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2024.3475271 · Cited by 10
Cohen, M. C., Kim, N., Ba, Y., Pan, A., Bhatti, S. A., Salehi, P., Sung, J., Blasch, E., Mancenido, M. V., & Chiou, E. K. (2025) · AI Magazine · doi:10.1002/aaai.70000 · Cited by 10
Roodposhti, M. S., Lucieer, A., Anees, A., & Bryan, B. A. (2019) · Remote Sensing · doi:10.3390/rs11172057 · Cited by 8
Artificial Intelligence: A New Tool in Oncologist's Armamentarium
Talwar, V., Chufal, K. S., & Joga, S. (2021) · Indian Journal of Medical and Paediatric Oncology · doi:10.1055/s-0041-1735577 · Cited by 8
Mehtab, F. H., & Mahmud, A. (2022) · Studies in computational intelligence · doi:10.1007/978-3-030-93921-2_22 · Cited by 7
Zhang, C., Correia, C., Weisskittel, T., Tan, S. H., Meng-Lin, K., Yu, G., Yao, J., Yeo, K. S., Zhu, S., Ung, C. Y., et al. (2022) · Frontiers in Immunology · doi:10.3389/fimmu.2022.920669 · Cited by 11
“Numerous efforts had been spent to develop explainable artificial intelligence (XAI) where the reasonings behind the predictions of AI algorithms can be traced and understood by humans (72, 73).”
BackgroundArtificial intelligence in bronchopulmonary dysplasia- current research and unexplored frontiers
Shah, M. U., Jain, D., Prasath, V. B. S., & Dufendach, K. R. (2022) · Pediatric Research · doi:10.1038/s41390-022-02387-z · Cited by 6
“34 Because of the “ black box ” nature of some of the AI algorithms, 35 particularly the deep learning-based algorithms, there may be clinician distrust if a result suggests an intervention that the clinician does not agree with.”
BackgroundKoidala, S. P., Manne, S. R., Ozimba, K., Rasheed, M. A., Bashar, S. B., Ibrahim, M. N., Selvam, A., Sahel, J., Chhablani, J. K., Jana, S., et al. (2023) · Scientific Reports · doi:10.1038/s41598-023-28512-4 · Cited by 7
SmokerViT: A Transformer-Based Method for Smoker Recognition
Khan, A. H., Khan, S., Hassan, B., Khan, R., & Zheng, Z. (2023) · Computers, materials & continua/Computers, materials & continua (Print) · doi:10.32604/cmc.2023.040251 · Cited by 10
Dutta, D. J., Sasaki, J., Bansal, A., Sugai, K., Yamashita, S., Li, G., Lazarski, C. A., Wang, L., Sasaki, T., Yamashita, C., et al. (2023) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2304074120 · Cited by 7
Artificial Intelligence in Radiology
Khan, M. R. (2024) · Elsevier eBooks · doi:10.1016/b978-0-323-87865-4.00011-x · Cited by 7
Two-step interpretable modeling of ICU-AIs
Lancia, G., Varkila, M. R. J., Cremer, O. L., & Spitoni, C. (2024) · Artificial Intelligence in Medicine · doi:10.1016/j.artmed.2024.102862 · Cited by 8
The UK Research Excellence Framework and the Matthew effect: Insights from machine learning
Balbuena, L. (2018) · PLoS ONE · doi:10.1371/journal.pone.0207919 · Cited by 11
Marginean, A. N., Groza, A. P., Nicoară, S., Muntean, G. A., Slăvescu, R. R., & Leţia, I. A. (2019) · IEEE International Conference on Intelligent Computer Communication and Processing (ICCP) · doi:10.1109/iccp48234.2019.8959714 · Cited by 7
Thought experiment: Decoding cognitive processes from the fMRI data of one individual
Wegrzyn, M., Aust, J., Barnstorf, L., Gippert, M., Harms, M., Hautum, A., Heidel, S., Herold, F., Hommel, S. M., Knigge, A.-K., et al. (2018) · PLoS ONE · doi:10.1371/journal.pone.0204338 · Cited by 6
Graells-Garrido, E., Peña-Araya, V., & Bravo, L. (2020) · Sustainability · doi:10.3390/su12156001 · Cited by 7
A Decentralized Approach to Explanatory Artificial Intelligence for Autonomic Systems
Houzé, É., Diaconescu, A., Dessalles, J.-L., Mengay, D., & Schumann, M. (2020) · IEEE International Conference on Autonomic Computing and Self-Organizing Systems Companion (ACSOS-C) · doi:10.1109/acsos-c51401.2020.00041 · Cited by 8
“However, more recent techniques, such as Deep Neural Networks [13], are typically “black box” models [3], with little to no room for understanding the reasons behind their decisions.”
BackgroundLearning like humans with Deep Symbolic NetworksPreprint
Zhang, Q., & Sornette, D. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1707.03377 · Cited by 8
“The Deep Symbolic Networks (DSN) model aims at solving the drawback of Deep learning that the obtained models remain black-boxes[3].”
BackgroundData Analytics and Deep Learning in Medicinal Chemistry
Bajorath, J. (2018) · Future Medicinal Chemistry · doi:10.4155/fmc-2018-0129 · Cited by 10
Machines are not moral role models
Cremer, D. D. (2022) · Nature Human Behaviour · doi:10.1038/s41562-022-01290-1 · Cited by 8
Mersha, M. A., Lamb, K., Wood, J., Alshami, A. M., & Kalita, J. K. (2024) · SSRN Electronic Journal · doi:10.2139/ssrn.4715286 · Cited by 7
Rosenbacke, R. (2024) · SSRN Electronic Journal · doi:10.2139/ssrn.4773350 · Cited by 9
Todino, M. D. (2025) · Encyclopedia · doi:10.3390/encyclopedia5010023 · Cited by 26
Explainable & Safe Artificial Intelligence in Radiology
Do, S. (2024) · Journal of the Korean Society of Radiology · doi:10.3348/jksr.2024.0118 · Cited by 7
Ma, W., Wang, Y., Ma, N., & Ding, Y. (2024) · Neuroscience · doi:10.1016/j.neuroscience.2024.12.045 · Cited by 7
Yang, L., Zhang, N., Jia, J., & Ma, Z. (2024) · Scientific Reports · doi:10.1038/s41598-024-83347-x · Cited by 6
Pacchioni, F., Germagnoli, G., Calbi, M., Agostoni, G., Sapienza, J., Repaci, F., D'Incalci, M. F., Spangaro, M., Cavallaro, R., & Bosia, M. (2025) · BioMedInformatics · doi:10.3390/biomedinformatics5010008 · Cited by 6
Cherubini, C., Cipriano, G., Saccotelli, L., Dimauro, G., Coppini, G., Carlucci, R., Fanizza, C., & Maglietta, R. (2025) · Ecological Informatics · doi:10.1016/j.ecoinf.2025.103066 · Cited by 10
An explainable artificial intelligence – human collaborative model for investigating patent novelty
Jang, H., & Yoon, B. (2025) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2025.110984 · Cited by 7
Hu, S., Wang, Y., Liu, J., & Yang, C. (2025) · Information Fusion · doi:10.1016/j.inffus.2025.103275 · Cited by 9
Degen, B. (2025) · Computers and Education Open · doi:10.1016/j.caeo.2025.100278 · Cited by 8
Weighted cycle-based identification of influential node groups in complex networks
Zheng, W., Shi, W., Fan, T., & Lü, L. (2025) · Physica A Statistical Mechanics and its Applications · doi:10.1016/j.physa.2025.130830 · Cited by 8
“Machine learning has emerged as a promising direction for addressing this challenge, but the “black-box” nature of many models [25] limits interpretability [26], constraining their practical application.”
Building Trustworthy AI: Transparency, Fairness, and Governance in the Digital Age
Liu, E. (2025) · AI & Future Society · doi:10.63802/afs.v1.i1.80 · Cited by 7
Chen, Y., & Chu, S. (2025) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw67362.2025.00336 · Cited by 7
“…target confirmation, decision validation, and consensus refinement, ensuring coordinated and re-silient operations while adapting to dynamic network conditions. making, often function as “black boxes,” obscuring the rationale behind their decisions and impeding reliability assessments [9, 56, 74].”
Contrastive explanations for machine learning predictions in chemistry
Lamens, A., & Bajorath, J. (2025) · Journal of Cheminformatics · doi:10.1186/s13321-025-01100-6 · Cited by 9
“Typically, contemporary ML models are “black boxes”, excluding the possibility to explain their decisions based on human considerations [1–3].”
Prospects and challenges of deep learning in gynecologic malignancies
Zhang, Y., & Zeng, L. (2025) · Frontiers in Oncology · doi:10.3389/fonc.2025.1592078 · Cited by 10
Rahman, M. A. (2026) · Intelligence-Based Medicine · doi:10.1016/j.ibmed.2026.100413 · Cited by 15
Graph-based unsupervised domain adaptation for fault diagnosis of HVAC systems
Ghalamsiah, N., Wen, J., Wu, T., Candan, K. S., & O’Neill, Z. D. (2025) · Building and Environment · doi:10.1016/j.buildenv.2025.114055 · Cited by 11
Zhou, Y., Wang, Y., & Su, J. (2025) · Scientific Reports · doi:10.1038/s41598-025-28532-2 · Cited by 9
Automation — Augmentation of Artificial Intelligence and Human Intelligence
Raj, R., Srivastava, A. K., & Behera, R. (2026) · Technovation · doi:10.1016/j.technovation.2025.103462 · Cited by 10
Verisimilar Percept Sequences Tests for Autonomous Driving Intelligent Agent Assessment
Watanabe, T., & Wolf, D. F. (2018) · Latin American Robotics Symposium, Brazilian Symposium on Robotics and Workshop on Robotics in Education (LARS/SBR/WRE) · doi:10.1109/lars/sbr/wre.2018.00048 · Cited by 8
Zipori, A. B., Kerley, C. I., Klein, A., & Kenney, R. (2022) · Journal of Neuro-Ophthalmology · doi:10.1097/wno.0000000000001682 · Cited by 6
“One of the major obstacles to clinical acceptance of AI models is what is often referred to as the “ black box ” of neural networks (30,31), which refers to the notion that the relationship between a model ’ s input and output is not fully understood, even by the creators of the model.”
BackgroundInteractive Collaborative Learning with Explainable Artificial Intelligence
Arnold, O., Golchert, S., Rennert, M., & Jantke, K. P. (2023) · Lecture notes in networks and systems · doi:10.1007/978-3-031-26876-2_2 · Cited by 10
YAPAY ZEKÂ BAĞLAMINDA OKUL LİDERLERİNİ BEKLEYEN ZORLUKLAR
Sincar, M. (2023) · Mustafa Kemal Üniversitesi Eğitim Fakültesi Dergisi · doi:10.56677/mkuefder.1407065 · Cited by 10
On the Explainability of Financial Robo-Advice Systems
Vilone, G., Sovrano, F., & Lognoul, M. (2024) · Communications in computer and information science · doi:10.1007/978-3-031-63803-9_12 · Cited by 9
Development of a multi-modal learning-based lymph node metastasis prediction model for lung cancer
Park, J.-M., Kim, S., Lim, J. H., Kim, C., You, S., Choi, J., Lim, J. H., Chang, J. W., Park, D., Lee, M.-W., et al. (2024) · Clinical Imaging · doi:10.1016/j.clinimag.2024.110254 · Cited by 6
The Digital Siren’s Call: Accepting Unethical AI Advice
Xu, C., Xu, H., Sun, Y., & Xiong, W. (2024) · International Journal of Human-Computer Interaction · doi:10.1080/10447318.2024.2400396 · Cited by 8
“While considerable research on AI ethics has centered on algorithmic fairness (Barocas & Selbst, 2016), transparency (Castelvecchi, 2016), and accountability (Mittelstadt et al., 2016), there remains a lacuna concerning human responses to unethical advice dispensed by AI systems.”
The role of parsimonious models in addressing mobility challenges
Barthélemy, M. (2024) · npj Sustainable Mobility and Transport · doi:10.1038/s44333-024-00011-7 · Cited by 11
“Although this is not always regarded as a bad thing [35], this lack of interpretability makes it challenging to understand how AI tools arrive at their decisions, hindering transparency and accountability in transportation systems.”
Metadata
- Authors
- Castelvecchi, D.
- Corresponding author
- Castelvecchi, Davide
- Venue
- Nature
- Published
- 2016-10
- Volume
- 538
- Issue
- 7623
- Pages
- 20-23
- ISSN
- 0028-0836 (ISSN-L)
- 1476-4687
- OpenAlex
- W2528491735
- Cited by
- 1,773
- References
- 6
- Citations by year
- 2026: 195 · 2025: 322 · 2024: 276 · 2023: 244 · 2022: 217 · 2021: 214 · 2020: 148 · 2019: 84 · 2018: 54 · 2017: 18 · 2016: 1
- FWCI
- 42.56
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- cc-by-sa
- www.springer.com/tdm (tdm)
- Locations
- Nature (publishedVersion, cc-by-sa)
- PubMed (publishedVersion)
- RePEc: Research Papers in Economics (submittedVersion)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Explainable Artificial Intelligence (XAI)
- Physical Sciences › Computer Science › Artificial Intelligence › Artificial Intelligence Applications
- Physical Sciences › Computer Science › Artificial Intelligence › Neural Networks and Applications
- Keywords
- artificial intelligence, machine learning, interpretability, explainability, trust in AI
- MeSH terms
- Machine Learning, Astronomy, Creativity, Humans, Physics, Robotics, Thinking, Neural Networks, Computer, Uncertainty, Datasets as Topic
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.
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
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.