Pith. sign in

REVIEW 4 major objections 4 minor 37 references

"It Warned Me Just at the Right Moment": Exploring LLM-based Real-time Detection of Phone Scams

T0 review · 4 major / 4 minor · reviewed 2026-08-09 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read An LLM that reads a phone call one utterance at a time flags scams with recall between 0.98 and 1.00, and an 'UNCERTAIN' option trades some of that recall for higher precision.

desk verdict A competent offline feasibility study of turn-by-turn LLM scam detection with a useful UNCERTAIN prompt, but the title's 'real-time' claim outruns what was actually measured. read the letter →

arxiv 2502.03964 v1 pith:XN63BTZY submitted 2025-02-06 cs.HC cs.CR

Zitong Shen , Sineng Yan , Youqian Zhang , Xiapu Luo , Grace Ngai , Eugene Yujun Fu This is my paper · ORCID
classification cs.HCcs.CR
keywords phonescamdetectionlargelanguagemodelsreal-timefraudconversationalAIuncertaintydeferralprecision-recalltrade-offChinesecalltranscripts
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Phone scams remain hard to stop because existing defenses act before the call (blacklists, caller ID) or after it (retrospective analysis), leaving the call itself unprotected. This paper argues that a large language model can fill that gap by reading an ongoing conversation one utterance at a time and classifying each new turn as fraudulent, safe, or uncertain. In experiments on Chinese transcripts, all five tested LLMs achieved recall between 0.98 and 1.00, meaning almost every scam call was flagged. The paper also shows that letting the model answer 'UNCERTAIN' raises precision on a synthetic dataset (from 0.70–0.77 to 0.80–0.85 for the weakest models) at a measurable cost in recall (0.90–0.98). The point of the work is that real-time, mid-call detection is within reach, with the main open problem being the trade-off between false alarms and alert timing.

What carries the argument

The load-bearing mechanism is turn-by-turn evaluation with conversation history: after each speaker's utterance, the full transcript so far is fed to an LLM with one of two prompts. The binary 'RT' prompt forces a FRAUD/SAFE decision; the 'UNC' prompt adds an UNCERTAIN category so the model can defer when evidence is thin. The deferral mechanism is what carries the precision improvement, because it prevents premature keyword-triggered alarms and lets later context resolve the call.

What would settle it

Run the same RT and UNC prompts on transcripts produced by a live automatic-speech-recognition pipeline, with imperfect transcription, overlapping speech, and realistic pauses, and compare precision and recall against Table 1; if recall drops materially or the alert arrives after the victim would have acted, the central claim fails.

Watch

Extended reading notes

Core claim

The paper's central claim is that turn-by-turn LLM analysis can detect phone scams in real time with high sensitivity: across GPT-4, GPT-4o, GLM4, Doubao-Pro-32k, and ERNIE-3.5-8k, the binary real-time prompt ('RT') scored recall between 0.98 and 1.00 on both authentic and synthetic Chinese call datasets, with several models at 1.00. The weakness of this real-time mode is precision: on the synthetic dataset, where benign calls deliberately contain suspicious keywords, RT precision ranges between 0.70 and 0.77 for most models, and the paper traces these false positives to the model deciding on partial context, e.g., flagging a flight rebooking call when the agent says 'payment.' Its proposed fix is an 'UNCERTAIN' option in the prompt ('UNC'), which lets the model defer judgment until more context arrives; this raises synthetic-dataset precision to 0.80–0.85 (and to 1.00 for one model) while recall drops to 0.90–0.98, and it visibly delays the alert in a worked scam-call example from utterance 6 to utterance 10. Against a retrospective baseline that sees the whole call, real-time detection trades precision for timeliness, and the paper argues this trade-off is the central design problem for any deployed system.

Load-bearing premise

The load-bearing assumption is that replaying clean, complete Chinese transcripts one utterance at a time behaves like detection on a live call; since no streaming speech-to-text, latency budget, or alert mechanism is built or measured, real-world transcription errors or delay could change the reported precision and recall.

Editorial extensions

If this is right

  • If the reported recall transfers to real calls, a detector could warn a user mid-conversation at the first clearly fraudulent utterance, before money or personal information changes hands.
  • Because the precision gap appears on synthetic keyword-rich calls, legitimate customer-service conversations about payments, IDs, or deadlines will be the main source of false alarms in deployment.
  • Choosing the UNCERTAIN option means accepting lower recall and later alerts in exchange for fewer false alarms; a deployed system must pick a point on this curve, not a single accuracy number.
  • The consistency across five different LLMs suggests that the result is not an artifact of one model family, so future improvements can be compared against roughly the same recall ceiling.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A natural extension the paper does not run is an end-to-end test with automatic speech recognition, speaker diarization, and network latency; measurement noise in those components could lower recall below the clean-transcript numbers.
  • The false-positive categories the paper identifies suggest a hybrid design: use UNCERTAIN deferral only when a call touches identity, payment, and urgency keywords, and alert immediately outside those contexts, preserving both precision and timeliness.
  • The same turn-by-turn protocol could be applied to non-Chinese languages, but the keyword-triggered false-alarm pattern means language-specific tuning would be needed before the precision numbers transfer.
  • The alert-timing trade-off could be formalized as a cost model: assign a dollar or trust cost to false alarms and a harm cost to missed or late scam alerts; the optimal deferral threshold would then follow from the user's or operator's loss function.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 4 minor

Summary. The paper proposes an LLM-based framework for detecting phone scams in real time. A system model with Scammer, User, and LLM-based Detector is presented, and the detection method classifies each conversational turn as FRAUD or SAFE, with an optional UNCERTAIN category to defer judgment. The evaluation uses Chinese transcripts from two datasets (authentic and synthetic) drawn from the authors' prior work [27] and five LLMs (GPT-4, GPT-4o, GLM4, Doubao-Pro-32k, ERNIE-3.5-8k). Table 1 reports accuracy, precision, recall, and F1 for binary real-time detection (RT), uncertainty-enabled detection (UNC), and a retrospective baseline (RET). The main empirical findings are that RT achieves recall between 0.98 and 1.00 across models and datasets, while precision on the synthetic dataset is lower (0.70–0.77); adding UNC raises synthetic precision to 0.83–0.85 at the cost of recall dropping to 0.90–0.98. The authors analyze false-positive categories, attribute them to limited context and keyword sensitivity, and discuss the precision–timeliness trade-off. The conclusion identifies notification strategies and user studies as future work.

Significance. If the reported results hold, the paper would show that LLMs can assess fraudulent intent from partial conversation context with high sensitivity, and that an uncertainty option is a workable mechanism for trading recall against precision. The contribution is potentially useful for the human-centered security community because it addresses an important harm and proposes a concrete detection approach. The paper has several strengths: it evaluates five diverse LLMs, uses two datasets, compares against a retrospective baseline, and includes concrete false-positive examples that illustrate the failure modes. However, the significance is conditional because the central 'real-time' claim rests on an offline simulation over clean transcripts, and the reported aggregate metrics do not establish that warnings arrive early enough to prevent harm.

major comments (4)
  1. [§4.1, Table 1] The unit of analysis for the metrics in Table 1 is not specified, and the surrounding text and examples indicate that a call is counted as detected if any single turn returns FRAUD. Under this 'ever flagged' aggregation, a recall of 1.00 is compatible with a system that only alerts on the final utterance of a scam call, which would not provide the timely intervention claimed in the title and RQ1. The paper must report per-utterance classification performance, the distribution of first-alert utterance indices, and the time-to-detection relative to call length; without these, the high recall values do not support the real-time warning claim.
  2. [§3.1.1, §2.2] The evaluation feeds clean written transcripts to the LLM one utterance at a time, with no streaming speech-to-text, no per-turn API latency, no turn-segmentation mechanism, and no alerting component. The real-time protocol is also underspecified: the number of prior utterances included in the context, the temperature, and the exact API versions are not reported. As a result, the precision and recall in Table 1 are upper-bound estimates for an offline simulation, not measurements of a real-time system. The authors should either implement and measure an end-to-end pipeline with ASR and latency, or explicitly reframe the contribution as turn-by-turn detection on transcripts and state that real-time performance remains untested.
  3. [§3.3, Table 1] Table 1 contains no sample sizes, confidence intervals, significance tests, or repeated runs. The differences between RT and UNC on the synthetic dataset (e.g., precision 0.70 to 0.83 for Doubao) are presented as definitive improvements, but with small datasets and no variance estimates these differences may not be reliable. The paper should report the number of scam and benign calls in each dataset and provide statistical tests or confidence intervals for the key precision and recall comparisons.
  4. [§4.3.2] The timeliness analysis is supported only by two anecdotal utterance indices from one example call: the RT system flags at Utterance 6 and UNC at Utterance 10. No distribution of detection latency is provided, and the paper does not report how often UNC delays detection or how large the delays are. Since the central value proposition is warning users before harm occurs, the paper needs a quantitative timeliness analysis, not just illustrative examples.
minor comments (4)
  1. [§1] The phrase 'combat of phone scams' should be 'combat phone scams'.
  2. [§4.2] The text says 'the detection method classifies the conversion as fraud'; 'conversion' should be 'conversation'.
  3. [§3.2] 'Generalize ability' should be 'generalization ability'.
  4. [§4.1, Table 1] The label 'Real-time Detection' for the RT and UNC columns is misleading because the experiments use offline transcripts; consider naming these columns 'Turn-by-turn (simulated)' or adding an explicit caveat in the table caption.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: the central RT/UNC results are empirical measurements, and the self-cited dataset and RET baseline are not used to derive the claims.

full rationale

The paper's central claim, that LLM turn-by-turn detection (RT) achieves 0.98-1.00 recall and that adding an UNCERTAIN option (UNC) trades recall for precision, is an empirical measurement over transcripts using five LLMs and two prompts. No parameter is fitted to the reported precision/recall numbers, and no equation defines the target metric in terms of an input quantity. The self-citations are limited to the dataset source [27] (Section 3.1) and the retrospective baseline RET [27] (Section 4.1). These citations are not load-bearing in a circular sense: the RT and UNC numbers in Table 1 are computed by the present paper's own experiment, and the RET baseline is a comparison point, not an input to the derivation. The RET baseline's lower recall and higher precision do not constrain the RT/UNC results by construction. The main substantive weakness is external validity, not circularity: the evaluation feeds clean transcripts (Section 3.1.1) and aggregates detection at call level, so 'real-time' performance on a live call is not directly measured. That is a correctness/validity concern, not a circularity concern, and per the instructions it is not scored under circularity. The paper is self-contained against the reported benchmark numbers, so the score is low.

Assumptions & free parameters 1 free parameters · 4 assumptions · 0 invented entities

The central claim rests on the validity of the transcript labels and on the assumption that offline turn-by-turn scoring approximates live detection. No numerical constants are fitted to data, but the unreported context-window length is a hand-chosen parameter that affects all reported numbers.

free parameters (1)
  • real-time context window (number of prior utterances) = not reported
    Section 2.2 says the detector uses 'prior conversational history' but never defines how many previous turns are included. This choice affects what the model sees and therefore all precision and recall numbers in Table 1.
assumptions (4)
  • domain assumption Authentic and synthetic transcript labels (scam vs. normal) are correct.
    Section 3.1 describes datasets sourced from YouTube and generated synthetically but gives no labeling protocol, inter-annotator agreement, or ground-truth validation. All metrics inherit these labels.
  • domain assumption Offline turn-by-turn scoring on clean transcripts approximates live real-time detection.
    Section 4.1 evaluates RT and UNC on pre-collected transcripts, not a deployed system with streaming ASR, so latency and transcription errors are excluded from the reported numbers.
  • domain assumption LLM API outputs are stable enough that single point estimates are meaningful.
    No temperature, seed, or repetition count is reported in Sections 3.2 or 4, so the results assume deterministic or near-deterministic responses.
  • domain assumption Results on Chinese-language calls generalize to other languages.
    All datasets are in Chinese (Section 3.1), but the conclusions are stated without restricting to Chinese-language calls. Cross-lingual scam phrasing and ASR behavior may differ.

how reviews work

0 comments
Cite this review

Pith. "Pith review of "It Warned Me Just at the Right Moment": Exploring LLM-based Real-time Detection of Phone Scams." pith.science (2026). https://pith.science/paper/XN63BTZY

@misc{pith2026250203964,
  author       = {Pith},
  title        = {Pith review of: "It Warned Me Just at the Right Moment": Exploring LLM-based Real-time Detection of Phone Scams},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XN63BTZY}},
  note         = {Machine review of arXiv:2502.03964}
}
read the original abstract

Despite living in the era of the internet, phone-based scams remain one of the most prevalent forms of scams. These scams aim to exploit victims for financial gain, causing both monetary losses and psychological distress. While governments, industries, and academia have actively introduced various countermeasures, scammers also continue to evolve their tactics, making phone scams a persistent threat. To combat these increasingly sophisticated scams, detection technologies must also advance. In this work, we propose a framework for modeling scam calls and introduce an LLM-based real-time detection approach, which assesses fraudulent intent in conversations, further providing immediate warnings to users to mitigate harm. Through experiments, we evaluate the method's performance and analyze key factors influencing its effectiveness. This analysis enables us to refine the method to improve precision while exploring the trade-off between recall and timeliness, paving the way for future directions in this critical area of research.

Figures

Figures reproduced from arXiv: 2502.03964 by the authors.

Figure 1
Figure 1. Taxonomy of Phone Scam Prevention Methods: A Temporal Perspective. [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. The system consists of three components, including “Scammer”, “User”, and “LLM-based Detector”. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Analysis of false positive scenarios in fraud detec [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Visualization of fraud detection scenarios in an ongoing normal phone call and a scam phone call. Suspicious patterns [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

37 extracted references · 29 canonical work pages

  1. [27]

    Zitong Shen, Kangzhong Wang, Youqian Zhang, Grace Ngai, and Eugene Y. Fu

    work page

  2. [1]

    Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Floren- cia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. 2023. Gpt-4 technical report. arXiv preprint arXiv:2303.08774 (2023)

    arXiv 2023

  3. [2]

    Baidu. 2024. ERNIE-3.5-8K. https://ai.baidu.com/ai-doc/WENXINWORKSHOP/ jlil56u11

    work page 2024

  4. [3]

    Jeremy Burke, Christine Kieffer, Gary Mottola, and Francisco Perez-Arce. 2022. Can Educational Interventions Reduce Susceptibility to Financial Fraud?. In Journal of Economic Behavior & Organization , Vol. 198. 250–266. doi:10.1016/j. jebo.2022.03.028

    doi:10.1016/j 2022

  5. [4]

    ByteDance. 2024. Doubao-Pro-32k. https://www.volcengine.com/product/ doubao

    work page 2024

  6. [5]

    Isha Chadalavada, Tianhui Huang, and Jessica Staddon. 2024. Distinguishing Scams and Fraud with Ensemble Learning. arXiv preprint arXiv:2412.08680 (2024). doi:10.48550/arXiv.2412.08680 arXiv:2412.08680 [cs.CR]

    work page Pith review arXiv doi:10.48550/arxiv.2412.08680 2024

  7. [6]

    Chen-Wei Chang, Shailik Sarkar, Shutonu Mitra, Qi Zhang, Hossein Salemi, He- mant Purohit, Fengxiu Zhang, Michin Hong, Jin-Hee Cho, and Chang-Tien Lu

    work page

  8. [7]

    Devon Cheyenne. 2024. Scammers can use AI Tools to Clone the Voices of You and Your Family—How to Protect Yourself.CNBC (Jan. 2024). https://www.cnbc.com/ 2024/01/24/how-to-protect-yourself-against-ai-voice-cloning-scams.html Re- trieved April 25, 2024

    work page 2024

Show all 37 references
  1. [8]

    Mitchell

    Marguerite DeLiema, Martha Deevy, Annamaria Lusardi, and Olivia S. Mitchell

  2. [9]

    Dzindolet, Scott A

    Mary T. Dzindolet, Scott A. Peterson, Regina A. Pomranky, Linda G. Pierce, and Hall P. Beck. 2003. The Role of Trust in Automation Reliance. International Journal of Human-Computer Studies 58, 6 (2003), 697–718. doi:10.1016/S1071- 5819(03)00038-7

    2003 doi

  3. [10]

    Team GLM, Aohan Zeng, Bin Xu, Bowen Wang, Chenhui Zhang, Da Yin, Dan Zhang, Diego Rojas, Guanyu Feng, Hanlin Zhao, et al. 2024. Chatglm: A fam- ily of large language models from glm-130b to glm-4 all tools. arXiv preprint arXiv:2406.12793 (2024)

    2024 arXiv

  4. [11]

    Gilad Gressel, Rahul Pankajakshan, and Yisroel Mirsky. 2024. Discussion Paper: Exploiting LLMs for Scam Automation: A Looming Threat. In Proceedings of the 3rd ACM Workshop on the Security Implications of Deepfakes and Cheapfakes (Singapore, Singapore) (WDC ’24). Association f...

    2024

  5. [12]

    Xinxin Hu, Hongchang Chen, Shuxin Liu, Haocong Jiang, Guanghan Chu, and Ran Li. 2022. BTG: A Bridge to Graph Machine Learning in Telecommunications Fraud Detection. Future Generation Computer Systems 137 (2022), 274–287. doi:10. 1016/j.future.2022.07.020

    2022

  6. [13]

    Xinxin Hu, Haotian Chen, Junjie Zhang, Hongchang Chen, Shuxin Liu, Xing Li, Yahui Wang, and Xiangyang Xue. 2024. GAT-COBO: Cost-Sensitive Graph Neural Network for Telecom Fraud Detection. IEEE Transactions on Big Data 10, 4 (2024), 528–542. doi:10.1109/TBDATA.2024.3352978

    2024

  7. [14]

    Aaron Hurst, Adam Lerer, Adam P Goucher, Adam Perelman, Aditya Ramesh, Aidan Clark, AJ Ostrow, Akila Welihinda, Alan Hayes, Alec Radford, et al. 2024. Gpt-4o system card. arXiv preprint arXiv:2410.21276 (2024)

    2024 arXiv

  8. [15]

    Liam Kennedy. 2014. AI Voice Scams are Coming: Here’s What You Need to Know. CHOICE Australia (2014). https://www.choice.com.au/consumers-and-data/ protecting-your-data/data-privacy-and-safety/articles/ai-voice-scams Retrieved April 25, 2024

    2014

  9. [16]

    Hong, and Jeffrey Nichols

    Tianshi Li, Julia Katherine Haines, Miguel Flores Ruiz De Eguino, Jason I. Hong, and Jeffrey Nichols. 2023. Alert Now or Never: Understanding and Predicting Notification Preferences of Smartphone Users. ACM Transactions on Computer- Human Interaction 29, 5, Article 39 (2023), ...

    2023 doi

  10. [17]

    Farhanim Mohamad Asri and Tengku Elena Tengku Mahamad. 2023. Anatomy of Phone Scams: Victims’ Recall on the Communication Phrases Used by Phone Scammers. In Proceedings of the International Conference on Communication and Media 2022 (Selangor, Malaysia) (i-COME ’22). Associati...

    2023 doi

  11. [18]

    Hossen Mustafa, Wenyuan Xu, Ahmad-Reza Sadeghi, and Steffen Schulz. 2018. End-to-End Detection of Caller ID Spoofing Attacks. In IEEE Transactions on Dependable and Secure Computing , Vol. 15. 423–436. doi:10.1109/TDSC.2016. 2580509

    2018 doi

  12. [19]

    Brian New, Lexi Salazar, Mike Lozano, and Scott Fralicks. 2024. Deepfakes of Elon Musk Are Contributing to Billions of Dollars in Fraud Losses in the U.S.CBS News Texas (24 Nov. 2024). https://www.cbsnews.com/texas/news/deepfakes-ai- fraud-elon-musk/ Accessed: 2024-01-22

    2024

  13. [20]

    Kenya Freeman Oduor and Eric N. Wiebe. 2008. The Effects of Automated Decision Algorithm Modality and Transparency on Reported Trust and Task Performance. In Proceedings of the Human Factors and Ergonomics Society Annual Meeting. 302–306. doi:10.1177/154193120805200422

    2008 doi

  14. [21]

    Sharbani Pandit, Koutilya Sarker, Roberto Perdisci, Mustaque Ahamad, and Duen Horng Yang. 2023. Combating Robocalls with Phone Virtual Assistant Mediated Interaction. In 32nd USENIX Security Symposium (USENIX Security 23) . USENIX Association, 463–479

    2023

  15. [22]

    Andrea Papenmeier, Dagmar Kern, Gwenn Englebienne, and Christin Seifert

  16. [23]

    If We Don’t Listen to Them, We Make Them Lose More than Money:

    Katalin Parti and Faika Tahir. 2023. “If We Don’t Listen to Them, We Make Them Lose More than Money:” Exploring Reasons for Underreporting and the Needs of Older Scam Victims. Social Sciences 12, 5 (2023), 264

    2023

  17. [24]

    Lu Peng and Rongheng Lin. 2018. Fraud Phone Calls Analysis Based on Label Propagation Community Detection Algorithm. In Proceedings of the 2018 IEEE World Congress on Services (SERVICES ’18) . 23–24. doi:10.1109/SERVICES.2018. 00025

    2018 doi

  18. [25]

    Lubna Razaq, Tallal Ahmad, Samia Ibtasam, Umer Ramzan, and Shrirang Mare

  19. [26]

    Sam Rogers. 2024. International Scammers Steal Over $1 Trillion in 12 Months in Global State of Scams Report 2024 . Global Anti-Scam Al- liance. https://www.gasa.org/post/global-state-of-scams-report-2024-1-trillion- stolen-in-12-months-gasa-feedzai Accessed: 2024-01-22

    2024

  20. [28]

    True Software Scandinavia AB. 2024. Truecaller. https://www.truecaller.com. Mobile application software

    2024

  21. [29]

    Tseng, Jia-Ching Ying, Che-Wei Huang, Yimin Kao, and Kuan-Ta Chen

    Vincent S. Tseng, Jia-Ching Ying, Che-Wei Huang, Yimin Kao, and Kuan-Ta Chen

  22. [30]

    Jian Xing, Miao Yu, Shupeng Wang, Yaru Zhang, and Yu Ding. 2020. Auto- mated Fraudulent Phone Call Recognition through Deep Learning. In Wireless Communications and Mobile Computing, Vol. 2020. 1–9. doi:10.1155/2020/8853468

    2020 doi

  23. [31]

    Kun Yu, Shlomo Berkovsky, Ronnie Taib, Dan Conway, Jianlong Zhou, and Fang Chen. 2017. User Trust Dynamics: An Investigation Driven by Differences in System Performance. In Proceedings of the 22nd International Conference on Intelligent User Interfaces. Association for Computi...

    2017

  24. [32]

    Combating Phone Scams with LLM-based Detection: Where Do We Stand? arXiv:2409.11643 [cs.CR] https://arxiv.org/abs/2409.11643

    arXiv

  25. [2015]

    In Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD ’15)

    FrauDetector: A Graph-Mining-based Framework for Fraudulent Phone Call Detection. In Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD ’15) . 2157–2166. doi:10.1145/ 2783258.2788623

  26. [2020]

    In The Journals of Gerontology: Series B, Vol

    Financial Fraud Among Older Americans: Evidence and Implications. In The Journals of Gerontology: Series B, Vol. 75. 861–868. doi:10.1093/geronb/gby151

    doi

  27. [2021]

    We Even Borrowed Money From Our Neighbor

    " We Even Borrowed Money From Our Neighbor" Understanding Mobile- based Frauds Through Victims’ Experiences. Proceedings of the ACM on human- computer interaction 5, CSCW1 (2021), 1–30

    2021

  28. [2022]

    ACM Transactions on Computer-Human Interaction 29, 4, Article 35 (March 2022), 33 pages

    It’s Complicated: The Relationship between User Trust, Model Accuracy and Explanations in AI. ACM Transactions on Computer-Human Interaction 29, 4, Article 35 (March 2022), 33 pages. doi:10.1145/3495013

    2022 doi

  29. [2024]

    arXiv preprint arXiv:2412.00621 (2024)

    Exposing LLM Vulnerabilities: Adversarial Scam Detection and Perfor- mance. arXiv preprint arXiv:2412.00621 (2024). doi:10.48550/arXiv.2412.00621 arXiv:2412.00621 [cs.CR]

    2024 doi

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.