TweepFake: About detecting deepfake tweets
Open accessThe DOI matches its Crossref record; 49 of the 50 references it cites were found in a registry or an index.
Abstract
The recent advances in language modeling significantly improved the generative capabilities of deep neural models: in 2019 OpenAI released GPT-2, a pre-trained language model that can autonomously generate coherent, non-trivial and human-like text samples. Since then, ever more powerful text generative models have been developed. Adversaries can exploit these tremendous generative capabilities to enhance social bots that will have the ability to write plausible deepfake messages, hoping to contaminate public debate. To prevent this, it is crucial to develop deepfake social media messages detection systems. However, to the best of our knowledge no one has ever addressed the detection of machine-generated texts on social networks like Twitter or Facebook. With the aim of helping the research in this detection field, we collected the first dataset of real deepfake tweets, TweepFake . It is real in the sense that each deepfake tweet was actually posted on Twitter. We collected tweets from a total of 23 bots, imitating 17 human accounts. The bots are based on various generation techniques, i.e., Markov Chains, RNN, RNN+Markov, LSTM, GPT-2. We also randomly selected tweets from the humans imitated by the bots to have an overall balanced dataset of 25,572 tweets (half human and half bots generated). The dataset is publicly available on Kaggle. Lastly, we evaluated 13 deepfake text detection methods (based on various state-of-the-art approaches) to both demonstrate the challenges that Tweepfake poses and create a solid baseline of detection techniques. We hope that TweepFake can offer the opportunity to tackle the deepfake detection on social media messages as well.
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
Figures


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
“Similarly, a text deepfake can be used to spread disinformation alongside manipulated images, amplifying its impact on social media platforms (Fagni et al., 2021).”
From Deepfake forensics: a survey of digital forensic methods for multimodal deepfake identification on social media (2024) · cites this paper as Fagni et al., 2021
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
“For instance, Fagni et al. (2021) highlighted the potential misuse of text-based deepfakes on Twitter but found existing detection methods to be inadequate.”
From Deepfake forensics: a survey of digital forensic methods for multimodal deepfake identification on social media (2024) · cites this paper as Fagni et al., 2021
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“The results demonstrate TweepFake’s challenges and provide a baseline for future deepfake detection techniques.”
From Deepfake forensics: a survey of digital forensic methods for multimodal deepfake identification on social media (2024) · cites this paper as Fagni et al. (2021)
The cited source says what the claim attributes to it.
In the citing paper
“According to their findings, a wide range of detectors based on ML or DL approaches and transformer-based utilizing transfer learning) had more trouble successfully identifying a deepfake tweet than a tweet that was actually produced by a human.”
From Deepfakes: current and future trends (2024) · cites this paper as Fagni et al. 2021
The cited source says what the claim attributes to it.
In the citing paper
“A noteworthy exception is the work by Fagni et al., who identify 23 self-disclosed bot accounts on Twitter and share the dataset (Fagni et al., 2021).”
From Anatomy of an AI-powered malicious social botnet (2024) · cites this paper as Fagni et al., 2021
3 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
Breiman, L. (2001) · Machine Learning · doi:10.1023/a:1010933404324 · Cited by 133,538 · cited 1×
“ed. Table 3. Description of the methods used in the experimentation. Encoding Method name Algorithm BoW+TF IDF log reg bow Logistic regression classifier [56] rand forest bow Random forest classifier [57] svc bow Support vector machine classifier [58] BERT log reg bert Logistic regression classifier rand forest bert R…”
MethodVaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L., & Polosukhin, I. (2017) · arXiv · doi:10.48550/arxiv.1706.03762 · Cited by 120,404 · cited 3×
“In 2017, the development of the selfattentionmechanism and the [25]’s transformer led to the improvement of the language models.”
BackgroundMethodInfluentialExploiting Generative AI to Scale up Intelligent Tutoring Systems
Jakubův, J., Chvalovský, K., Goertzel, Z., Kaliszyk, C., Olšák, M., Piotrowski, B., Schulz, S., Suda, M., & Urban, J. (2023) · DROPS (Schloss Dagstuhl – Leibniz Center for Informatics) · doi:10.4230/lipics.itp.2023.19 · Cited by 76,658
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2018) · arXiv · doi:10.48550/arxiv.1810.04805 · Cited by 73,173 · cited 4×
“and perform training on a specific dataset (typically small) for very few epochs [27].”
BackgroundMethodInfluentialGradient-based learning applied to document recognition
Lecun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998) · Proceedings of the IEEE · doi:10.1109/5.726791 · Cited by 58,805
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2019) · Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers) · doi:10.18653/v1/n19-1423 · Cited by 54,325 · cited 4×
“and perform training on a specific dataset (typically small) for very few epochs [27].”
BackgroundMethodInfluentialLearning Phrase Representations using RNN Encoder–Decoder for Statistical Machine Translation
Cho, K., van Merrienboer, B., Gulcehre, C., Bahdanau, D., Bougares, F., Schwenk, H., & Bengio, Y. (2014) · Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.3115/v1/d14-1179 · Cited by 25,545 · cited 1×
“Characters char cnn Single CNN network [59] using internal char embeddings representation char gru Single GRU network [60] using internal char embeddings representation char cnngru Combined CNN and GRU networks using internal char embeddings representation”
MethodRoBERTa: A Robustly Optimized BERT Pretraining Approach
Liu, Y., Ott, M., Goyal, N., Du, J., Joshi, M., Chen, D., Levy, O., Lewis, M., Zettlemoyer, L., & Stoyanov, V. (2019) · arXiv · doi:10.48550/arxiv.1907.11692 · Cited by 16,515 · cited 5×
“However, OpenAI [31] proved it wrong on GPT-2 generated texts: they showed that fine-tuning a RoBERTa-based detector achieved consistently higher accuracy than fine-tuning a GPT-2 based detector with equivalent capacity.”
MethodInfluentialAdvances in neural information processing systems 7
(1996) · Computers & Mathematics with Applications · doi:10.1016/s0898-1221(96)90463-0 · Cited by 14,281
Transformers: State-of-the-Art Natural Language Processing
Wolf, T., Debut, L., Sanh, V., Chaumond, J., Delangue, C., Moi, A., Cistac, P., Rault, T., Louf, R., Funtowicz, M., et al. (2020) · Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing: System Demonstrations · doi:10.18653/v1/2020.emnlp-demos.6 · Cited by 8,741
Research works printed but not linked: 1
- Fagni T. Scripts used in “TweepFake: about Detecting Deepfake Tweets”; 2021. Repository: GitHub [Internet]. Available from: https://github.com/tizfa/tweepfake_deepfake_text_detection [cited 2021 March 16]. PMC8118345 33984021
Cited by
Items 101–196 of 196
The potential of GPTs for enhanced information access and user services at academic libraries
Hamad, F. F., & Shehata, A. M. K. (2024) · IFLA Journal · doi:10.1177/03400352241298958 · Cited by 13
“The GPT models demonstrated substantial improvements in text generation and language modeling tasks, creating opportunities for their implementation in various fi elds, including library sciences (Fagni et al., 2021).”
Lost in Transformation: Rediscovering LLM-Generated Campaigns in Social Media
Britta, G., Pohl, J., Winkelmann, H., Stampe, L., & Grimme, C. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-47896-3_6 · Cited by 15
Detecting and Understanding Textual Deepfakes in Online Reviews
Kowalczyk, P., Röder, M., Dürr, A., & Thiesse, F. (2022) · Proceedings of the Annual Hawaii International Conference on System Sciences · doi:10.24251/hicss.2022.184 · Cited by 9
“To identify deepfake tweets created with GPT-2 Fagni et al. [31] use LR, SVM and RF.”
Ruffin, M., Seo, H., Xiong, A., & Wang, G. (2024) · Proceedings of the International AAAI Conference on Web and Social Media · doi:10.1609/icwsm.v18i1.31392 · Cited by 15
“Deepfake is a combined word of “deep learning” and “fake” (Mirsky and Lee 2021), which refers to synthetic media (e.g., images (Wang et al. 2020), text (Fagni et al. 2020), and videos (Lyu 2020)) forged by deep learning methods.”
BackgroundA Comparison of DeepSeek and other LLMs
Gao, T., Jin, J., Ke, Z. T., & Moryoussef, G. (2026) · The American Statistician · doi:10.1080/00031305.2025.2611010 · Cited by 71
“Note that aside from LLMs, we may apply other algorithms to this problem (Solaiman et al. 2019; Zellers et al. 2019; Gehrmann, Strobelt, and Rush 2019; Ippolito et al. 2020; Adelani et al. 2020; Fagni et al. 2021; Mitchell et al. 2023; Kashtan and Kipnis 2024).”
AI-Generated Text Detection: Challenges and Future Directions
An, B. (2023) · International Journal of Asian Language Processing · doi:10.1142/s2717554523300025 · Cited by 8
AIRABIC: Arabic Dataset for Performance Evaluation of AI Detectors
Alshammari, H., & EI-Sayed, A. (2023) · International Conference on Machine Learning and Applications · doi:10.1109/icmla58977.2023.00127 · Cited by 19
Performance Analysis of Deepfake Text Detection Techniques on Social-media
Gaur, A., Singh, S. K., & Saxena, P. (2024) · International Conference on Distributed Computing and Optimization Techniques (ICDCOT) · doi:10.1109/icdcot61034.2024.10515626 · Cited by 10
“The study uses the TweepFake dataset which is a publicly accessible labeled dataset having human-generated and bot-generated tweets on X [9].”
MethodCoAT: Corpus of artificial texts
Shamardina, T., Saidov, M., Fenogenova, A., Tumanov, A., Zemlyakova, A., Lebedeva, A., Gryaznova, E., Shavrina, T., Mikhailov, V., & Artemova, E. (2024) · Natural language processing. · doi:10.1017/nlp.2024.38 · Cited by 11
The harm of deepfakes: a scoping review of deepfakes’ negative effects on human mind and behavior
Diel, A., Lalgi, T. J., Mellis, F. S., Teufel, A., & Bäuerle, A. (2025) · AI & Society · doi:10.1007/s00146-025-02774-0 · Cited by 25
“Deepfakes have found various forms of use and misuse in the marketplace, such as in the form of AI-generated advertising (Chen 2024; Kietzmann et al. (Kietzmann, et al., 2021)) or fake reviews (Fagni et al. 2021).”
MARKMyWORDS: Analyzing and Evaluating Language Model Watermarks
Piet, J., Sitawarin, C., Fang, V., Mu, N., & Wagner, D. (2025) · IEEE Conference on Secure and Trustworthy Machine Learning (SaTML) · doi:10.1109/satml64287.2025.00012 · Cited by 103
“Many schemes can be found in the literature, including OpenAI’s classifiers [24, 25], GLTR [26], DetectGPT [27], and others [28, 29, 30, 31, 32].”
Ibrahim, K. H. S., Al Otaibi, D., & Sibai, F. N. (2025) · International Journal of Computer-Assisted Language Learning and Teaching · doi:10.4018/ijcallt.375428 · Cited by 5
A novel approach to identify deepfake text using social media data
Devi, B. T., & Rajasekaran, R. (2025) · Social Network Analysis and Mining · doi:10.1007/s13278-025-01536-6 · Cited by 6
“Studies have shown that around 9 to 20% of Twitter users are automated accounts, also known as bots (Fagni et al. 2021).”
InfluentialA Step Towards Modern Disinformation Detection: Novel Methods for Detecting LLM-Generated Text
Nathanson, S., Yoo, Y., Na, D., Cao, Y., & Watkins, L. A. (2024) · IEEE Military Communications Conference (MILCOM) · doi:10.1109/milcom61039.2024.10773838 · Cited by 13
Leveraging Deep Learning Models for Machine-Generated Text Detection Using Transformer-Based Models
Suthanthiradevi, P., Revathy, G., Rejini, K., & Lakshmi.V, M. (2024) · International Conference on Innovative Computing, Intelligent Communication and Smart Electrical Systems (ICSES) · doi:10.1109/icses63760.2024.10910837 · Cited by 6
“Researchers employed a tweet dataset from several bots, including cyborg, social, spam, and sock puppet [11].”
Looking for Traces of Textual Deepfakes in Bulgarian on Social Media
Temnikova, I., Marinova, I., Gargova, S., Margova, R., & Koychev, I. K. (2023) · International conference Recent advances in natural language processing · doi:10.26615/978-954-452-092-2_122 · Cited by 17
“Although there are generators for different languages, the existing datasets for detectors training are mostly in English (Fagni et al., 2021; Liyan-age et al., 2022), with Chinese (Chen et al., 2022a) and Russian 12 (Posokhov et al., 2022; Shamardina et al., 2022) also available.”
BackgroundInfluentialLeveraging Multi-Model Linguistic Fusion for Enhanced AI Text Generation Evasion
Khan, A. A., & Rajagopalan, S. (2025) · International Generative AI and Computational Language Modelling Conference (GACLM) · doi:10.1109/gaclm67198.2025.11231821 · Cited by 18
Exploring the Capability of ChatGPT to Reproduce Human Labels for Social Computing Tasks
Zhu, Y., Zhang, P., Haq, E.-U., Hui, P., & Tyson, G. (2025) · Lecture notes in computer science · doi:10.1007/978-3-031-78548-1_2 · Cited by 18
GigaCheck: Detecting LLM-generated ContentPreprint
Tolstykh, I., Tsybina, A., Yakubson, S., Gordeev, A., Dokholyan, V., & Kuprashevich, M. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5381590 · Cited by 6
Cheng, Z., Zhou, L., Jiang, F., Wang, B., & Li, H. (2025) · Proceedings of the ACM on Web Conference 2025 · doi:10.1145/3696410.3714770 · Cited by 3
“(iii) Supervised fine-tuning classifiers, trained on annotated data [3, 26, 45], have shown effectiveness in detecting LLM-generated text across domains, such as news [29, 61], social media (e.g., Twitter) [19, 43], and academic papers [59].”
Echoes Without Integration: Strategic Resonance and the Limits of Radical Right Transnationalism
Zanotti, L., Machuca, F. V., & Duque, F. R. (2025) · Política y Sociedad · doi:10.5209/poso.99134 · Cited by 5
AI-Generated Text Detection and Source Identification
Tatavarthi, A. P., Abri, F., & Attar, N. (2025) · Journal of Advances in Information Technology · doi:10.12720/jait.16.7.1030-1041 · Cited by 3
Exploring Naive Approaches to Tell Apart LLMs Productions from Human-written Text
Giudice, O., Maggi, A., & Nardelli, M. (2023) · International Conference on Natural Language Processing and Information Retrieval (NLPIR) · doi:10.1145/3639233.3639354 · Cited by 4
“The most popular approach for designing machine-generated text detectors rely on fine-tuned LMs , which exploit transfer learning to retrain the LM to detect generated text (e.g., [7, 11, 13, 18, 22]).”
BackgroundA Metric-Based Detection System for Large Language Model Texts
Le, L., & Tran, D. T. (2025) · ACM Transactions on Management Information Systems · doi:10.1145/3704739 · Cited by 8
“Traditionally, the typical works in detection of LLM texts train supervised detection models on top of deep embedding representations [2, 9, 17, 43].”
A feature selection-aided deep learning based deepfake video detection method
Mohiuddin, S., Roy, A., Pani, S., Malakar, S., & Sarkar, R. (2025) · Multimedia Tools and Applications · doi:10.1007/s11042-025-20877-0 · Cited by 5
Rajagopal, T., Chandrashekaran, V., & Ilango, V. (2023) · Journal of Communication and Management · doi:10.58966/jcm2023243 · Cited by 5
How Natural Language Processing Enables AIGC Recognition? --Latest Trends and Future Prospects
Wang, W., & Qiao, H. (2024) · International Conference on Software Engineering and Information Management (ICSIM) · doi:10.1145/3647722.3647738 · Cited by 4
“CHEAT: The CHEAT dataset [31] stands as one of the most comprehensive public resources, focusing on detecting artificially generated deceptive academic content from ChatGPT.”
BackgroundUnmasking the AI Hand: A Machine Learning Approach to Deciphering Authorship
Bhandarkar, A., M, M. A. D., Vishwachetan, D., Mushtaq, A., Kadam, D., & Saxena, S. (2024) · International Conference for Innovation in Technology (INOCON) · doi:10.1109/inocon60754.2024.10512010 · Cited by 3
“Tiziano Fagni et al. [4] proposed a method that helps detect Deepfake tweets using TF-IDF vectorizer and Bag of Words.”
MethodBeyond Words: Stylometric Analysis for Detecting AI Manipulation on Social Media
Ullah, U., Laudanna, S., Vinod, P., Sorbo, A. D., Visaggio, C. A., & Canfora, G. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-70879-4_11 · Cited by 6
Using OpenAI models as a new tool for text analysis in political leaders’ unstructured discoursePreprint
Bianchini, G. E., Zanotti, L., & Meléndez, C. (2023) · PsyArXiv (OSF Preprints) · doi:10.31234/osf.io/kdngb · Cited by 4
Forging the Forger: An Attempt to Improve Authorship Verification via Data Augmentation
Corbara, S., & Moreo, A. (2024) · IEEE Access · doi:10.1109/access.2024.3481161 · Cited by 4
“Moreover, due to the recent significant advances in language modelling, powerful Neural Networks (NNs) are now able to autonomously generate ‘‘coherent, non-trivial and human-like’’ text samples [13, p. 1], that can be exploited as fake news or propaganda [13], [14].”
Öztürk, M. (2024) · Akdeniz Üniversitesi İletişim Fakültesi Dergisi · doi:10.31123/akil.1538165 · Cited by 3
Yadav, A., & Shunmuga Priya, M. C. (2025) · Operations Research Forum · doi:10.1007/s43069-025-00486-1 · Cited by 5
Tulga, A. Y. (2025) · EDPACS · doi:10.1080/07366981.2025.2564783 · Cited by 3
“In this context, numerous academic studies, particularly in computer science and engineering, have focused on the development of deepfake detection algorithms (Fagni et al., 2021; Kumar et al., 2025).”
Sohail, M., Hongying, Z., Abdullah, M., Caraffini, F., Javed, A., & Eshkiki, H. (2025) · Knowledge-Based Systems · doi:10.1016/j.knosys.2025.115206 · Cited by 3
Dual-branch approach for AI-generated scientific content detection
Alhijawi, B., Jarrar, R., Awajan, S., & AbuAlRub, A. (2026) · Knowledge-Based Systems · doi:10.1016/j.knosys.2026.115440 · Cited by 2
Social Bots Spreading Disinformation About Finance: Research Trends, and Ethical Challenges
Pohl, J., Griesbach, M., samiei, A., & de Araujo, A. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-52082-2_6 · Cited by 2
Development of a Neural Network Model for Recognizing Russian-Language Generated Texts
Nikolaev, K. S. (2024) · International Multi-Conference on Engineering, Computer and Information Sciences (SIBIRCON) · doi:10.1109/sibircon63777.2024.10758447 · Cited by 4
A Survey on Writing Style Change Detection: Current Literature and Future Directions
Hashemi, A., & Shi, W. (2025) · Machine Intelligence Research · doi:10.1007/s11633-025-1544-6 · Cited by 3
Detection and Analysis of AI-Generated Malicious Content
Arafah, M. E., Al-Banna, A.-K., & Aladawi, A. (2025) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-3373-0832-6.ch011 · Cited by 4
Khalid, M., Mushtaq, M. F., Akram, U., Safran, M., Alfarhood, S., & Ashraf, I. (2025) · Scientific Reports · doi:10.1038/s41598-025-10661-3 · Cited by 7
“It utilized various deep learning techniques, including LSTM, RNN, and GPT-2, to analyze results, with the RoBERTa-based technique outperforming others by achieving a 90% accuracy score 6 .”
Agarwal, A., & Uzair, M. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-78498-9_5 · Cited by 1
Detection of AI-Generated Text
Bharadwaj, R., Dugad, R., Gile, A., Patil, G., & Hatyalikar, S. (2024) · International Conference on Progressive Innovations in Intelligent Systems and Data Science (ICPIDS) · doi:10.1109/icpids65698.2024.00032 · Cited by 2
“[11 ] Figure 7: The presented model for detecting bot-generated text content.”
Master of Deceit: Comparative Analysis of Human and Machine-Generated Deceptive Text
Trinh, Q. M., Zarin, S., & Rezapour, R. (2025) · ACM Web Science Conference (WebSci) · doi:10.1145/3717867.3717914 · Cited by 6
“This may influence perceived authenticity [19].”
Meng, D., Wang, Z., Phuntsho, T., & Gonsalves, T. (2025) · Egyptian Informatics Journal · doi:10.1016/j.eij.2025.100752 · Cited by 1
Deepfake Text Detection Using NLP and ML
Akashdeep, Mitra, H., & Gashroo, O. B. (2025) · International Conference on Computing, Intelligence, and Application (CIACON) · doi:10.1109/ciacon65473.2025.11189404 · Cited by 1
Alharthi, R. S., Ojo, S., Nathaniel, T. I., Samee, N. M. A., Umer, M., Jamjoom, M. M., Alsubai, S., & Khan, J. (2025) · International Journal of Computational Intelligence Systems · doi:10.1007/s44196-025-01025-w · Cited by 5
RepreGuard: Detecting LLM-Generated Text by Revealing Hidden Representation Patterns
Chen, X., Wu, J., Yang, S., Zhan, R., Wu, Z., Luo, Z., Wang, D., Yang, M., Chao, L. S., & Wong, D. F. (2025) · Transactions of the Association for Computational Linguistics · doi:10.1162/tacl.a.61 · Cited by 30
DeBERTa-based framework for detecting machine-generated content on social media: a comparative study
Sweidan, S., Farouk, N., Abouhawwash, M., Askar, S., & Taha, M. (2026) · Journal Of Big Data · doi:10.1186/s40537-025-01349-6 · Cited by 3
Deepfake Detection Using Deep Learning and Fast Text Embeddings
Bunyan, V. J., Nethala, T. R., RamaKrishna, B., Gelli, A., & Praveen, C. (2026) · Lecture notes in electrical engineering · doi:10.1007/978-981-96-7991-1_59 · Cited by 1
Sadiq, S., Ullah, S., Abuzinadah, N. E., Alharthi, R. S., Alabdullah, B. I., Umer, M., Alsubai, S., & Kryvinska, N. (2026) · International Journal of Computational Intelligence Systems · doi:10.1007/s44196-026-01192-4 · Cited by 2
“This research employs the TweepFake dataset [48], which includes 25,572 tweets collected from 40 accounts 17 belonging to humans and 23 to bots.”
Using AI to control AI-assisted plagiarism in L2 writing: a critical review of the literature
Ibrahim, K. H. S. (2026) · Language Testing in Asia · doi:10.1186/s40468-026-00433-9 · Cited by 2
Deepfake on Social Media: Harnessing Deep Learning for Identifying Falsified Tweet
Kaur, J. K. (2024) · INTERANTIONAL JOURNAL OF SCIENTIFIC RESEARCH IN ENGINEERING AND MANAGEMENT · doi:10.55041/ijsrem32156 · Cited by 2
Transformer-Based Models for Detecting Bots on Twitter
Sallah, A., Alaoui, E. A. A., & Agoujil, S. (2024) · Springer proceedings in energy · doi:10.1007/978-3-031-57022-3_16 · Cited by 1
Making Original 2D Visual Deepfakes With a Generative AI Tool
Hai-Jew, S. (2024) · Advances in multimedia and interactive technologies book series · doi:10.4018/979-8-3693-2927-6.ch006 · Cited by 1
Big data visualization through the lens of Peirce’s visual sign theory
Friedman, A., & Thellefsen, M. (2022) · Punctum International Journal of Semiotics · doi:10.18680/hss.2022.0007 · Cited by 3
Sadiq, S., & Ullah, S. (2023) · SSRN Electronic Journal · doi:10.2139/ssrn.4494619 · Cited by 1
تأثير الفضاء السيبراني على البيئة الثقافية والتنشئة الاجتماعية والعلاقات الأسرية
قطر, أ. د. ذ. م. د. ا. ا. د., & العشاري, ف. (2022) · Academic Journal of Research and Scientific Publishing · doi:10.52132/ajrsp/v4.44.4
Adversarial Attacks Against Detecting Bot Generated Text
Jun, L. E. H. (2023) · Springer eBooks · doi:10.1007/978-981-19-7222-5_37
Uma Abordagem a DeepFake via Algoritmos de Aprendizagem Profunda
Rodrigues, G. S., & Silva, C. A. (2023) · doi:10.5753/encompif.2023.230761
“Na literatura existem diversos casos destes tipos de manipulações visuais em cenários polı́ticos [Appel and Prietzel 2022], de redes sociais [Fagni et al. 2021], de entretenimento [Usukhbayar and Homer 2020] entre outros.”
BackgroundPaperMill Detection in Scientific Content
Razis, G., Anagnostopoulos, K., Metaxas, O., Stefanidis, S.-D., Zhou, H. G., & Anagnostopoulos, I. E. (2023) · International Workshop on Semantic and Social Media Adaptation and Personalization (SMAP) · doi:10.1109/smap59435.2023.10255173 · Cited by 2
“The work in [15] relies on the pretrained RoBERTa model for detecting human- from AIgenerated tweets, by outperforming traditional ML-based and Deep Neural Network models.”
MethodInaugural Defense and Security Research Symposium of the Purdue Military Research Institute
Laux, C. M., Dietz, J. E., Rapp, R. R., & Robert, L. L. (2024) · Purdue University Press eBooks · doi:10.5703/1288284317732
On Text Granularity and Metric Frameworks for Large Language Model Content Detection
Le, L., & Tran, D. T. (2024) · IEEE International Conference on Big Data (BigData) · doi:10.1109/bigdata62323.2024.10825505
Toward Robust Generative AI Text Detection: Generalizable Neural Model
Abburi, H., Pudota, N., Veeramani, B., Bowen, E., & Bhattacharya, S. (2024) · International Conference on Machine Learning and Applications · doi:10.1109/icmla61862.2024.00255 · Cited by 4
“The solutions capable of distinguishing AI-generated text from human-authored text roughly fall into three categories [8]: simple statistical techniques, zero-shot detection, and fine-tuning based detection.”
On the Gap Between AI-Generated and Human-Written Patent Texts
Xiao, Z., Hu, W. P., Wu, Y., Chen, W., Li, H., & Liu, X. (2025) · Lecture notes in computer science · doi:10.1007/978-981-96-2882-7_14 · Cited by 1
Enhancing the robustness of Fast-DetectGPT against paraphrase attacks
Li, S. (2024) · International Conference on Computers and Artificial Intelligence Technology (CAIT) · doi:10.1109/cait64506.2024.10962860 · Cited by 3
Deepfake Text Detection Using Deep Convolutional Neural Networks with Explainable AiPreprint
Sadiq, S., & Ullah, S. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5265122 · Cited by 1
Amplifying Training Data Exposure Through Fine-Tuning With Pseudo-Labeled Memberships
Oh, M. G., Ahn, H. E., Park, L. H., & Kwon, T. (2025) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2025.3613882 · Cited by 2
“Among the extensive prior research addressing this issue [21], [28], [80], we focus on a popular zero-shot machine-generated text detection strategy called DetectGPT [44].”
Towards Robust Detection of AI-Generated Text and Image Using Language and Vision Neural Models
Baskota, R., Dahal, D., Karki, Y., Acharya, S., & Acharya, P. (2025) · European Journal of Applied Science Engineering and Technology · doi:10.59324/ejaset.2025.3(6).01
Feng, Y., Wang, H., Li, J., Cao, Z., & Yan, L. (2025) · Systems · doi:10.3390/systems13110990 · Cited by 1
“However, it has emerged that some strategies, such as substituting synonyms for original characters, can alter the statistical characteristics of the generated text, making the detection task more difficult [12].”
InfluentialWhen Less is More: Minimal Prompts with LoRA for LLM Text Detection
Wang, S., Fang, R., Li, M., He, Z., & Song, S. (2025) · Lecture notes in computer science · doi:10.1007/978-981-95-3352-7_22 · Cited by 1
An exploratory study on LLM-generated code and comments in code repositories
Ji, Y., Wang, J., Zhou, Y. X., Chen, F., & Yang, H. (2026) · Journal of Systems and Software · doi:10.1016/j.jss.2026.113026 · Cited by 1
“For example, Fagni et al. [51] trained several supervised models to classify LLM-generated content on social media platforms.”
Detection of AI-Generated texts with BERT variants
Balara, V., & Machová, K. (2025) · International Conference on Emerging eLearning Technologies and Applications (ICETA) · doi:10.1109/iceta67772.2025.11280096
“• TweepFake [24] : a labelled dataset containing human and machine generated tweets.”
Seeing Isn't Believing: Addressing the Societal Impact of Deepfakes in Low-Tech Environments
Wasi, A. T., Priti, R. N., Khan, M. A., Rahman, A., & Islam, M. R. (2025) · doi:10.1145/3746275.3762206 · Cited by 4
“Deepfakes, a form of AI-generated multimedia content, encompass images, videos, audio, and text designed to convincingly mimic real media [16].”
A Comparative Analysis of Transformer Models in Social Bot Detection
Veit, R., & Lones, M. A. (2026) · Advances in intelligent systems and computing · doi:10.1007/978-3-032-07938-1_30
MAGENTA+: Detection of Arabic Machine-Generated Text Across Domains
Houmaine, A., & Rosso, P. (2026) · Lecture notes in networks and systems · doi:10.1007/978-3-032-03695-7_41
MAF-Detect: A Multi-Scale Adaptive Fusion Framework for Zero-Shot Detection of LLM-Generated Text
Bai, Z., & Zheng, Y. L. (2025) · International Conference on Machine Learning, Big Data and Business Intelligence (MLBDBI) · doi:10.1109/mlbdbi67855.2025.11331454 · Cited by 1
“Fagni et al. (2019) proposed a detection method based on LLM generated text, which can detect high-quality text generated by LLM. Fagni T et al.[1] found that fine-tuning RoBERTa can achieve excellent classification results in various settings.”
Rath, S., Rautaray, S. S., & Pandey, M. (2025) · IEEE International Conference on Applied Electromagnetics, Signal Processing, & Communication (AESPC) · doi:10.1109/aespc67542.2025.11326844
El corpus ROBOT-TALK para el reconocimiento del origen robótico de textos en español
Simón, L. A., & Cesteros, A. M. F. (2026) · Alfinge Revista de Filología · doi:10.21071/arf.v37i.18687 · Cited by 3
A Training-free Method for LLM Text AttributionPreprint
Radvand, T., Abdolmaleki, M., Mostagir, M., & Tewari, A. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6169527 · Cited by 7
Elgammal, Z., & Alhajj, R. (2026) · Social science computer review · doi:10.1177/08944393261432625
Détection des éléments générés automatiquement Contenu académique
Liyanage, V. P. (2024) · theses.fr (ABES) · doi:10.70675/80dff7a0z6e58z4151zb20cz6f0bc537ce85
Detecting AI-Generated Text: A Survey
Alaktif, A., Chergui, M., Ammoumou, A., & Gmira, F. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/407609
Ng, K. C., Ke, P. F., So, M. K. P., & Tam, K. Y. (2026) · ACM Transactions on Management Information Systems · doi:10.1145/3816728
Out of Order: On the Importance of Word Positions in Explaining Text Classification
Stockdiek, J. L., Britta, G., Griesbach, M., & Grimme, C. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-21477-5_29
Robust Detection of AI-Generated Text: Insights on Evolving LLMs and Adversarial Data
Guo, Y., Wang, B., Jiang, Z., Tian, S., Wang, K., Ke, W., & Luo, H. (2026) · Communications in computer and information science · doi:10.1007/978-981-95-8525-0_18
Ramesh, C., Reddy, A. R. M., & Noorullah, R. (2025) · International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-SMAC) · doi:10.1109/i-smac65734.2025.11597862
A Transformer-Based Model for Detection of AI-Generated Fake News: A Multi-perspective Exploration
Qin, H., Wu, H., Zhang, D. H., & Chen, L. (2026) · Lecture notes in computer science · doi:10.1007/978-981-92-3435-6_10
Statistical and Machine Learning Methods for Detecting AI-Generated Content in Gen AI Systems
Kumar, K. P., Soumya, E., & Mani, J. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-2600-0368-8.ch008
A Novel Hybrid Architecture for Deepfake Text Detection on Social Media
Lakshmi, M. A., Sumithra, B., Srija, A., Malathi, A., & Venkat, D. (2026) · International Conference on Computer Networks, Big Data and IoT (ICCBI) · doi:10.1109/iccbi68589.2026.11619670
“This diversity is essential for addressing the cross-generator generalization challenges highlighted in prior studies [7], [8].”
InfluentialJournalism in the Shadows: AI Deepfakes, Credibility, and the Fragility of Media Trust
Okocha, M. U. (2026) · Journalism Practice · doi:10.1080/17512786.2026.2719552 · Cited by 1
Combining probabilistic features and semantic features for AI-Generated text detection
Yu, Y., & Gao, W. (2026) · International Journal of Scientific Research and Management (IJSRM) · doi:10.18535/ijsrm/v14i09.ec01
Meng, D., & Gonsalves, T. (2026) · Neural Networks · doi:10.1016/j.neunet.2026.109562
Simplified Classification of Human- and LLM-authored Tweets Using Text Features
Dunn, M. H. (2026) · IEEE National Aerospace and Electronics Conference (NAECON) · doi:10.1109/naecon70028.2026.11675018
“Previous work by Fagni et al. [11] achieved some success in domain-specific detection of fake content on Twitter.”
InfluentialCharacterizing an LLM-driven Social Network: The Case of Chirper.ai
Zhu, Y., He, Y., Haq, E.-U., Tyson, G., & Hui, P. (2026) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3816910 · Cited by 23
“To assess the generalizability of our fine-tuned models in classifying AI-generated text across different platforms, we also evaluate their performance on three other well-established datasets: • TweepFake [19] : The TweepFake dataset is a collection of tweets designed for detecting deepfake tweets.”
Metadata
- Authors
- Fagni, T., Falchi, F., Gambini, M., Martella, A., & Tesconi, M.
- Affiliations
- Institute of Informatics and Telematics (IT)
- Istituto di Scienza e Tecnologie dell'Informazione "Alessandro Faedo" (IT)
- University of Pisa (IT)
- University of Trento (IT)
- Corresponding author
- Falchi, Fabrizio
- Venue
- PLOS ONE
- Publisher
- Public Library of Science (PLoS)
- Published
- 2021-05-13
- Volume
- 16
- Issue
- 5
- Pages
- e0251415
- ISSN
- 1932-6203 (ISSN-L)
- PMC
- PMC8118345
- OpenAlex
- W3046357466
- Cited by
- 185
- References
- 50
- Citations by year
- 2026: 24 · 2025: 45 · 2024: 50 · 2023: 40 · 2022: 18 · 2021: 7
- FWCI
- 39.82
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- PLoS ONE (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- CINECA IRIS Institutial research information system (University of Pisa) (submittedVersion)
- Institutional Research Information System University of Turin (University of Turin) (submittedVersion, cc-by)
- RePEc: Research Papers in Economics (submittedVersion)
- ISTI Open Portal (submittedVersion, other-oa)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Zenodo (CERN European Organization for Nuclear Research) (submittedVersion, cc-by)
- arXiv (Cornell University)
- Funders
- Horizon 2020 Framework Programme (871042, 95191)
- Topics
- Social Sciences › Sociology and Political Science › Misinformation and Its Impacts
- Physical Sciences › Computer Science › Information Systems › Spam and Phishing Detection
- Physical Sciences › Computer Science › Artificial Intelligence › Authorship Attribution and Profiling
- Keywords
- deepfake detection, AI-generated text detection, social bots, Twitter, text generation models, GPT-2, Markov chains, recurrent neural networks, LSTM, bot detection
- MeSH terms
- Artificial Intelligence, Humans, Language, Markov Chains, Social Media
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.