Synthetic data could be better than real data
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
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?
Because synthetic data are similar to but not the same as real data, they can be used to eliminate or mask personal identifiers and protect the confidentiality of human participants.
“because synthetic data are similar to but not the same as real data, they can be used to eliminate or mask personal identifiers and protect the confidentiality of human participants [31, 81, 200].”
From The ethics of using artificial intelligence in scientific research: new guidance needed for a new tool (2025) · cites this paper as [200]
This unethical use of AI is becoming increasingly likely since generative models can be used to create synthetic datasets from scratch or make alternative versions of existing datasets.
“This unethical use of AI is becoming increasingly likely since generative models can be used to create synthetic datasets from scratch or make alternative versions of existing datasets [50, 155, 200, 214].”
From The ethics of using artificial intelligence in scientific research: new guidance needed for a new tool (2025) · cites this paper as [200]
Cites
Dropout Prediction in MOOCs: Using Deep Learning for Personalized Intervention
Xing, W., & Du, D. (2018) · Journal of Educational Computing Research · doi:10.1177/0735633118757015 · Cited by 324 · see citing works
2016 IEEE International Conference on Data Science and Advanced Analytics
(2016) · IEEE Computational Intelligence Magazine · doi:10.1109/mci.2016.2545818 · Cited by 191 · see citing works
Flagged first, then by citations
1 of the 3 references couldn’t be matched to a known paper.
Cited by
Sorted by citations
Resnik, D. B., & Hosseini, M. (2025) · AI and Ethics · doi:10.1007/s43681-024-00493-8 · Cited by 228 · see citing works
“Also, because synthetic data are similar to but not the same as real data, they can be used to eliminate or mask personal identifiers and protect the confidentiality of human participants [31, 81, 200].”
From Fiction to Fact: The Growing Role of Generative AI in Business and Finance
Chen, B., Wu, Z., & Zhao, R. (2023) · Journal of Chinese Economic and Business Studies · doi:10.1080/14765284.2023.2245279 · Cited by 190 · see citing works
Kryshtafovych, A., Schwede, T., Topf, M., Fidelis, K., & Moult, J. (2023) · Proteins: Structure, Function, and Bioinformatics · doi:10.1002/prot.26617 · Cited by 148 · see citing works
“Figure S1 shows that in CASP15 there is a tendency for performance to fall-off with decreasing alignment depth, and this is much more pronounced when the ratio of alignment depth to target length is below 0.1 (less than 10 appropriately diverse sequences per 100 residues).”
BackgroundInfluentialSynthetic data in biomedicine via generative artificial intelligence
van Breugel, B., Liu, T., Oglić, D., & van der Schaar, M. (2024) · Nature Reviews Bioengineering · doi:10.1038/s44222-024-00245-7 · Cited by 97 · see citing works
“Beyond imitating real data, synthetic data can also solve some of the limitations of using real data 27–32 .”
MDD-ShipNet: Math-Data Integrated Defogging for Fog-Occlusion Ship Detection
Wang, N., Wang, Y., Feng, Y., & Wei, Y. (2024) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2024.3394573 · Cited by 52 · see citing works
“Moreover, researchers have found that models trained on synthetic data perform even better than those trained on real data in videos with fewer background objects 3 [52].”
BackgroundGenerating and evaluating synthetic data in digital pathology through diffusion models
Pozzi, M., Noei, S., Robbi, E., Cima, L., Moroni, M., Munari, E., Torresani, E., & Jurman, G. (2024) · Scientific Reports · doi:10.1038/s41598-024-79602-w · Cited by 51 · see citing works
Understanding Human-AI Workflows for Generating Personas
Shin, J., Hedderich, M. A., Rey, B. J., Lucero, A., & Oulasvirta, A. (2024) · Designing Interactive Systems Conference · doi:10.1145/3643834.3660729 · Cited by 38 · see citing works
“Often, they yielded incomplete personas because of truncation of a few sentences on account of GPT-4’s token limitation 2 .”
BackgroundSynthetic Health Data: Real Ethical Promise and Peril
Susser, D., Schiff, D. S., Gerke, S., Cabrera, L. Y., Cohen, I. G., Doerr, M. J., Harrod, J. B., Kostick, K. M., McNealy, J. E., Meyer, M. N., et al. (2024) · The Hastings Center Report · doi:10.1002/hast.4911 · Cited by 25 · see citing works
Joffe, I., Felobes, G., Elgouhari, Y., Talebi-Kalaleh, M., Mei, Q., & Chui, Y. H. (2025) · Journal of Computing in Civil Engineering · doi:10.1061/jccee5.cpeng-6037 · Cited by 25 · see citing works
Generative AI and the Workforce: What Are the Risks?Preprint
Walkowiak, E., & MacDonald, T. (2023) · SSRN Electronic Journal · doi:10.2139/ssrn.4568684 · Cited by 23 · see citing works
Marchesi, R., Micheletti, N., I-Hsien Kuo, N., Barbieri, S., Jurman, G., & Osmani, V. (2025) · PLOS Computational Biology · doi:10.1371/journal.pcbi.1013080 · Cited by 29 · see citing works
“GANs have shown incredible results in generating realistic images [18], text [19], and speech [20] in addition to improving privacy [21].”
Liu, K., Jin, X., Cheng, S., Gao, S., Yin, L., & Lu, F. (2023) · International Journal of Geographical Information Systems · doi:10.1080/13658816.2023.2292570 · Cited by 21 · see citing works
“In light of the above problems, generating synthetic trajectories that are statistically close to the real data without recycling real information is a potential solution (Pappalardo et al. 2023, Savage 2023).”
BackgroundSynthetic pre-training for neural-network interatomic potentials
Gardner, J. L. A., Baker, K. T., & Deringer, V. L. (2023) · Machine Learning Science and Technology · doi:10.1088/2632-2153/ad1626 · Cited by 20 · see citing works
Simulation and the reality gap: Moments in a prehistory of synthetic data
Steinhoff, J., & Hind, S. (2025) · Big Data & Society · doi:10.1177/20539517241309884 · Cited by 19 · see citing works
“Despite being in its infancy, synthetic data is cast by proponents as the cure for nearly all the problems associated with the machine learning approach to arti fi cial intelligence (AI), from labor costs to privacy and bias concerns (Savage, 2023).”
A Deep Learning-Based Pipeline for the Generation of Synthetic Tabular Data
Panfilo, D., Boudewijn, A., Saccani, S., Coser, A., Svara, B., Chauvenet, C. R., Mami, C. A., & Medvet, E. (2023) · IEEE Access · doi:10.1109/access.2023.3288336 · Cited by 16 · see citing works
“Deep learning (DL)-based generative models provide a reliable alternative to data anonymization [9].”
BackgroundAustin, J. A., Lobo, E. H., Samadbeik, M., Engstrom, T., Philip, R., Pole, J. D., & Sullivan, C. M. (2024) · Journal of Medical Internet Research · doi:10.2196/58637 · Cited by 15 · see citing works
“Such methods include common data models (CDMs) [16], federated learning (FL) [17], and synthetic data generation [18] supported by strong cross-sector collaboration.”
Romano, A. (2024) · Digital Geography and Society · doi:10.1016/j.diggeo.2024.100108 · Cited by 12 · see citing works
The ontological politics of synthetic data: Normalities, outliers, and intersectional hallucinations
Lee, F., Hajisharif, S., & Johnson, E. (2025) · Big Data & Society · doi:10.1177/20539517251318289 · Cited by 12 · see citing works
“Synthetic data are often presented as solving “ external ” data use problems — like privacy and regulation — speci fi c-ally because it they are not real data (Jacobsen, 2023; Offenhuber, 2024; Savage, 2023).”
Morgan, B., Murali, A. R., Preston, G., Sima, Y. A., Chamorro, L. A. M., Bourantas, C. V., Torii, R., Mathur, A., Baumbach, A., Jacob, M. C., et al. (2023) · Frontiers in Cardiovascular Medicine · doi:10.3389/fcvm.2023.1221541 · Cited by 11 · see citing works
“Synthetic manipulations have recently been introduced to Machine Learning to overcome the excessive requirement of well annotated data for AI algorithms (13).”
BackgroundInfluentialLi, P., Zhang, H., Li, W., Huang, D., Guo, Z., Chen, J., Zhang, J., Song, X., Zhao, P., Yan, J., et al. (2025) · Computers Environment and Urban Systems · doi:10.1016/j.compenvurbsys.2025.102252 · Cited by 10 · see citing works
Defect detection in photolithographic patterns using deep learning models trained on synthetic data
Shinde, P. P., Pai, P. P., Adiga, S. P., Mayya, K. S., Seo, Y., Hwang, M., Go, H., & Park, C.-M. (2025) · Heliyon · doi:10.1016/j.heliyon.2025.e43377 · Cited by 10 · see citing works
“Synthetically trained models can be as good as models trained on real data 41 .”
Fischer, A., Cameron, G. J., Tilus, C., Espey, J., & Badiee, S. (2025) · Data & Policy · doi:10.1017/dap.2025.10015 · Cited by 10 · see citing works
“Second, it is not only the supply but also the analytic methods that need clarification and standardization to leverage data value, particularly when using uncertainty science and synthetic data generation (Andree et al., 2023; Savage, 2023).”
Malthouse, E. C., Masłowska, E., Strycharz, J., Block, M. P., & Araujo, T. (2024) · Journal of Advertising · doi:10.1080/00913367.2024.2403609 · Cited by 9 · see citing works
“Models trained in this way can then create new datasets containing different data points than the original ones but retaining the relationships of interest to the researcher (Savage 2023).”
Song, J., Ashktorab, Z., Pan, Q., Dugan, C., Geyer, W., & Malone, T. W. (2025) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3757486 · Cited by 9 · see citing works
“This approach using machine-generated synthetic data aligns with the broader research goal of improving privacy and enhancing the representativeness of study results, as discussed in [50].”
Synthetic data and ELSI-focused computational checklists—A survey of biomedical professionals’ views
Wagner, J. K., Cabrera, L. Y., Gerke, S., & Susser, D. (2024) · PLOS Digital Health · doi:10.1371/journal.pdig.0000666 · Cited by 7 · see citing works
“While synthetic data might currently be a “relatively niche pursuit” [36], the funding opportunities to accelerate AI/ML generally and synthetic data specifically (including but not limited to its use in digital twins and the beginnings of data scientists creating longitudinal synthetic health…”
Critical Assessment of Methods of Protein Structure Prediction (CASP) – Round XV
Kryshtafovych, A., Schwede, T., Topf, M., Fidelis, K., & Moult, J. (2023) · doi:10.22541/au.169658651.11658182/v1 · Cited by 5 · see citing works
High-Fidelity Synthetic Face Generation for Rosacea Skin Condition from Limited Data
Mohanty, A., Sutherland, A., Bezbradica, M., & Javidnia, H. (2024) · Electronics · doi:10.3390/electronics13020395 · Cited by 5 · see citing works
Digital Disruption in Art: A Comprehensive Analysis of AI and NFT Market Dynamics
Włodarczyk, M. (2024) · Annales Universitatis Mariae Curie-Skłodowska sectio H Oeconomia · doi:10.17951/h.2024.58.2.171-193 · Cited by 3 · see citing works
Trustworthy Learning Analytics for Smart Learning Ecosystems
Svetec, B., & Divjak, B. (2025) · Interaction design & architecture(s)/ID&A Interaction design & architecture(s) · doi:10.55612/s-5002-064-002sp · Cited by 3 · see citing works
Once promised, forever believed? An essay on the blind spots of AI promises
Krokowski, T., & Hirsch‐Kreinsen, H. (2025) · AI & Society · doi:10.1007/s00146-025-02765-1 · Cited by 3 · see citing works
Tempered enthusiasm by interviewed experts for synthetic data and ELSI checklists for AI in medicine
Cabrera, L. Y., Wagner, J. K., Gerke, S., & Susser, D. (2025) · AI and Ethics · doi:10.1007/s43681-024-00652-x · Cited by 3 · see citing works
“In the biomedical context, several have pointed to the potential of synthetic data to protect privacy, a key consideration when sharing health data [ 13, 17, 35 ].”
Shan, Z., Yang, F., Shi, X., & Cui, Y. (2025) · ISPRS International Journal of Geo-Information · doi:10.3390/ijgi14050182 · Cited by 3 · see citing works
Ramzan, F., Kiberu, Y., Jathanna, N., Jamil‐Copley, S., Clayton, R. H., & Chen, C. (2025) · Lecture notes in computer science · doi:10.1007/978-3-032-00652-3_20 · Cited by 3 · see citing works
Generación masiva de historias clínicas sintéticas con ChatGPT: un ejemplo en fractura de cadera
Calvo-Lorenzo, I., & Llano, I. U. (2024) · Medicina Clínica · doi:10.1016/j.medcli.2023.11.027 · Cited by 5 · see citing works
On the Conceptualization of Synthetic Data: Why Not All Synthetic Data are Useful and Some are
Pashevich, E., & Hansen, J. U. (2026) · Minds and Machines · doi:10.1007/s11023-026-09781-0 · Cited by 2 · see citing works
“Utility can be tested via comparing predictions made by the model trained on synthetic data with the predictions made by a model trained on an original dataset (Savage, 2023).”
Massive generation of synthetic medical records with ChatGPT: An example in hip fractures
Calvo-Lorenzo, I., & Llano, I. U. (2024) · Medicina Clínica (English Edition) · doi:10.1016/j.medcle.2023.11.033 · Cited by 2 · see citing works
Yao, Y., Jiang, Y., Yu, Q., Yuan, J., Li, J., Xu, J., Liu, S., & Zhang, H. (2024) · ISPRS International Journal of Geo-Information · doi:10.3390/ijgi13070261 · Cited by 2 · see citing works
“Recent research found that synthetic data could be better than real data, as real data come with issues beyond privacy considerations, such as being expensive to produce and maintain, emphasizing the potential of machine-generated datasets in protecting privacy and addressing data skewness [8].”
BackgroundAlgorithms, Bytes, and Chips: The Emerging Political Economy of Foundation ModelsPreprint
Mills, S. (2024) · SSRN Electronic Journal · doi:10.2139/ssrn.4834417 · Cited by 2 · see citing works
Traveller: Travel-pattern aware trajectory generation via autoregressive diffusion models
Luo, Y., Zhang, S., Liu, K., Xu, Y., & Yin, L. (2025) · Information Fusion · doi:10.1016/j.inffus.2025.103766 · Cited by 2 · see citing works
Anti-Regulatory AI: How "AI Safety" is Leveraged Against Regulatory Oversight
Yew, R.-J., & Judge, B. (2025) · ACM Conference on Equity and Access in Algorithms, Mechanisms, and Optimization (EAAMO) · doi:10.1145/3757887.3763017 · Cited by 1 · see citing works
Digital phantoms in medical research: Synthetic data and the pursuit of ground truth
Högberg, C., & Winter, P. (2026) · Big Data & Society · doi:10.1177/20539517261447840 · Cited by 1 · see citing works
“An article in Nature , ‘ Synthetic data could be better than real data …’ (Savage, 2023), captures this shift, noting that barriers to real-world data access are pushing researchers towards synthetic equivalents.”
High Fidelity Synthetic Face Generation for Rosacea Skin Condition from Limited DataPreprint
Mohanty, A., Sutherland, A., Bezbradica, M., & Javidnia, H. (2023) · Research Square · doi:10.21203/rs.3.rs-2987104/v1 · Cited by 1 · see citing works
Braghin, C., & Cremonini, M. (2024) · Elsevier eBooks · doi:10.1016/b978-0-443-13223-0.00053-9
Traditional Research versus Advanced Algorithms: “Is Survey in the Last Days?”
Gabriel, M. L. D. D. S., Mazzon, J. A., Isabella, G., Limongi, R., Lopes, E. L., & Brei, V. A. (2024) · Revista de Administração Contemporânea · doi:10.1590/1982-7849rac2024240246.en
Pesquisa Tradicional versus Algoritmos Avançados: “O Survey Está Com os Dias Contados?”
Gabriel, M. L. D. D. S., Mazzon, J. A., Isabella, G., Limongi, R., Lopes, E. L., & Brei, V. A. (2024) · Revista de Administração Contemporânea · doi:10.1590/1982-7849rac2024240246.por
Artificially Generated Minorities (AGMs): The Veneer of Algorithmic Bias Correction
Jaja, I. R. (2025) · Communications in computer and information science · doi:10.1007/978-3-032-11733-5_34
Lu, Y., Liu, G., Li, X., & Jin, Z. (2026) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2026.113759
Saxena, S. (2026) · Lecture notes in networks and systems · doi:10.1007/978-981-96-8987-3_43
Improving rTMS Response Prediction in Depression Using Augmented fMRI Connectivity Data
Sharma, S. (2026) · IEEE International Conference on Interdisciplinary Approaches in Technology and Management for Social Innovation (IATMSI) · doi:10.1109/iatmsi68868.2026.11465482
“Utilizing synthetic data presents a promising approach to address this limitation [18], which can help create more diverse datasets [29].”
Synthetic Medical Data vs. The Four Principles: Ethical Trade-Offs and Zombie Data
Friedman, Y., & Smajdor, A. (2026) · Health Care Analysis · doi:10.1007/s10728-026-00596-9
- Authors
- Savage, N.
- Corresponding author
- Savage, Neil
- ORCID
- Savage, Neil
- Venue
- Nature
- Published
- 2023-04-27
- ISSN
- 0028-0836 (ISSN-L)
- 1476-4687
- OpenAlex
- W4367176123
- Cited by
- 53
- References
- 3
- Citations by year
- 2026: 7 · 2025: 18 · 2024: 19 · 2023: 9
- FWCI
- 4.57
- Type
- journal article
- Language
- en
- Open access
- closed
- Licence
- Locations
- Nature (publishedVersion)
- PubMed (publishedVersion)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Privacy-Preserving Technologies in Data
- Social Sciences › Safety Research › Ethics and Social Impacts of AI
- Social Sciences › Decision Sciences › Management Science and Operations Research › demographic modeling and climate adaptation
- Keywords
- synthetic data, machine-generated data, data privacy, data accuracy, synthetic data quality
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
