Pith. sign in

REVIEW 3 major objections 4 minor 1 cited by

Optimizing AI-Assisted Code Generation

T0 review · 3 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash

Pith's one-line read The paper argues that an encoder-decoder model that analyzes code for security before generating it, then retrains on its own labeled outputs, could make AI code generation secure and usable for non-experts.

desk verdict A decent, well-cited survey of security issues in AI code generation, wrapped in an unvalidated framework that the authors themselves admit still needs testing. read the letter →

arxiv 2412.10953 v1 pith:6FCRTB7L submitted 2024-12-14 cs.SE cs.AIcs.LG

classification cs.SEcs.AIcs.LG
keywords AIcodegenerationLLMsecurityencoder-decoderarchitecturevulnerabilitydetectionnon-expertprogrammingself-trainingfeedbacklooppromptengineeringforGood
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

The paper proposes that AI code generators can be made safe enough for non-experts if the generator is built as an encoder-decoder model that first analyzes code for security and then generates code meeting those standards. It surveys current LLM-based code generators and finds that they produce vulnerable code at scale and that novice users cannot reliably evaluate that code. The paper then assembles a conceptual architecture: validated training data, static and dynamic analysis, sandboxed execution, and a feedback loop that labels generated code and feeds it back into training. If this architecture works, it would give people without programming expertise a tool that produces secure, executable conventional code and AI models.

What carries the argument

The load-bearing mechanism is the encoder-decoder model's two-stage design, paired with an automated feedback loop that labels generated code as secure or insecure using static analysis tools, sandbox execution, and the LLM's own judgment. The encoder produces a security-aware representation of the input; the decoder generates code conditioned on that representation; and the loop re-integrates correctly labeled outputs into the training data.

What would settle it

Run the proposed labeling pipeline on a held-out benchmark of security-critical programming scenarios with expert-annotated ground truth; if the pipeline's secure/insecure labels disagree with the experts on a substantial share of vulnerable samples, or if a model trained with the loop still produces vulnerable code at a rate comparable to the untrained base model, the central claim fails.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central claim is that an encoder-decoder (sequence-to-sequence) architecture is the right foundation for a secure code-generation platform, because the encoder can analyze and classify code for security while the decoder generates code that meets the learned security standards. The paper proposes a zero-trust training pipeline in which datasets are validated with security analysis tools, generated code is executed in a sandbox, and the success or error signals are looped back into training as labeled examples. It also argues that continuous runtime feedback, prompt support, and prefix-based hardening methods allow the model to improve during operation and adapt to non-expert users.

Load-bearing premise

The whole design rests on the assumption that the automated feedback loop can reliably label generated code as secure or insecure; if the labels are wrong, feeding them back into training would reinforce vulnerabilities instead of removing them.

Editorial extensions

If this is right

  • A single AI platform using this design could in principle generate secure executable code and AI models without requiring users to write or review code.
  • The iterative labeling loop could turn raw code sources into high-quality labeled corpora for security training, addressing the current scarcity of labeled security data.
  • Prompt support and prefix-based hardening techniques could decouple output security from the user's language proficiency, reducing the novice-user vulnerability identified in user studies.
  • The zero-trust data validation concept could be reused for other LLM-based tools where training-data poisoning is a concern.

Reading between the lines

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

  • The paper leaves unstated that the strongest test of its design is measuring the precision and recall of its automated labels against expert-annotated code, and then the trained model's vulnerability rate on held-out security scenarios.
  • If the feedback loop's labels are noisy, self-training could amplify rather than remove vulnerabilities; comparing the loop's labels with independent static analyzers on the same code would bound that risk.
  • The same architecture could be evaluated for the AI-model-generation use case by checking whether models generated by the platform themselves produce secure code, which would close the loop recursively.
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

3 major / 4 minor

Summary. The manuscript surveys recent work on AI-assisted code generation, security challenges, attacks and defenses, and then sketches a conceptual platform for generating secure, efficient, functional code and AI models accessible to non-experts. Sections 2–6 review tool capabilities, user studies, vulnerabilities, prompt injection/jailbreaks, and mitigation strategies; Section 7 assembles these into a proposed encoder-decoder-based system with secure data curation, self-training feedback, prompt support, and AI-model generation. The stated aim is to develop such a tool, but the paper presents no implementation or empirical evaluation; Section 7.9 defers effectiveness testing to future hackathons.

Significance. If validated, the proposed platform would address a genuine need: democratizing secure AI/code generation for non-experts while mitigating known security risks. The survey is broad and cites many recent empirical results, such as DeVAIC's 94% accuracy on incomplete Python code, SVEN's security-rate improvement from 59.1% to 92.3%, and Perry et al.'s user study, and it summarizes real attacks and defenses in a structured way. However, the central contribution is a design sketch: no code, data, or measurements support the claimed benefits of the self-training loop, and the encoder-decoder architectural choice is asserted rather than demonstrated. The work is best read as a research agenda or vision paper, not a validated engineering contribution. The strengths are the breadth of the survey and the explicit attention to non-expert users and AI-for-good applications.

major comments (3)
  1. [§7.3 (and Figure 5)] The continuous self-training loop assumes that automatically generated security labels from static analyzers, sandbox execution, and the LLM itself are accurate enough to improve the generator. The paper provides no evidence for this accuracy and even notes in Section 4.2 that most static analysis tools cannot handle incomplete code; DeVAIC is cited as the only tool for incomplete Python code, with a benchmark accuracy of 94% on 35 CWEs. That benchmark does not measure precision and recall on the generated, often non-executable code inside the loop, and no comparison against a control trained on verified gold labels is offered. If the labeling function has systematic errors, self-training could amplify vulnerabilities rather than remove them; the paper must either supply such measurements or explicitly reframe this as an open research risk instead of a design guarantee.
  2. [§7.1] The choice of an encoder-decoder architecture is load-bearing for the proposed two-stage analysis-and-generation workflow, but the justification is an assertion that this architecture is 'particularly well suited' rather than a comparative argument. The empirical studies cited earlier in the paper (Sections 4.2 and 4.4) found that decoder-based Codex achieved the highest code generation capability and security repair performance, while encoder-based CodeBERT was weaker on security measures. This undercuts the unstated assumption that combining both roles in one encoder-decoder model is optimal. The authors should either cite direct evidence for encoder-decoder superiority on security-tuned code generation or present the architecture as one option among several, with explicit trade-offs.
  3. [§7.9 (and Introduction)] The stated aim of the paper is to develop a tool that generates secure, efficient, and functional code and AI models for non-experts, but Section 7.9 says only that effectiveness 'should be tested through hackathons' as suggested by [86], with no metrics, baselines, or acceptance criteria. The central claim is therefore unsupported by any implementation or evaluation. This can be fixed by reframing the paper as a research agenda that identifies open problems and a proposed approach, with hackathon studies as future work; as written, the conclusion overstates what has been established.
minor comments (4)
  1. [Section 7] Placeholder text 'für anderes paper' appears immediately before the Section 7 heading; this leftover note must be removed.
  2. [References] Reference entries are incomplete and inconsistent: many lack publication years (e.g., [3], [4], [6], [9], [24], [50]), some duplicate the same work under different numbers ([9] and [10]), and several entries mix German and English metadata (e.g., [17], [48], [100]).
  3. [Section 4.2] The sentence beginning '10In addition' appears to contain a typo, and the spelling of 'CodeX' vs. 'Codex' is inconsistent across Table 1 and the running text.
  4. [§7.2] The sandbox execution step is described as feeding errors back into training, but the paper does not specify which test oracles decide whether generated code is 'functional'; this is closely related to Major Comment 1 and should be clarified.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper is a survey and conceptual proposal with no derivation, fitted parameters, or load-bearing self-citations.

full rationale

Score 0. The paper is a literature survey plus a proposed conceptual framework (Section 7). It contains no equations, no fitted parameters, no numerical predictions, and no claim that a result follows from first principles. The Section 7.1 choice of encoder-decoder architecture is a design rationale ('the encoder enables the analysis and classification of the code in terms of its security. Then the decoder ensures that the generated code meets the required security standards'), not a derivation, and it is supported by an external citation [94]. The feedback loop described in Sections 7.2 and 7.3 is presented as a proposal ('we propose', 'the resulting code will be executed in a sandboxed environment'), with no empirical assertion that the loop has already been shown to improve security; the paper explicitly acknowledges the limitation that most static analysis tools cannot handle incomplete code [11]. The reference list contains no works by the present authors, so there is no self-citation chain and no imported uniqueness or ansatz. The internal placeholder note 'für anderes paper' before Section 7 suggests the section was drafted with possible reuse in mind, but this does not make any argument circular. The paper's weakness is that its central claims are unvalidated proposals, which is a correctness and evidence concern, not circularity.

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

The central framework depends on four domain assumptions that are asserted rather than tested. No free parameters or invented entities are introduced; the proposal is an assembly of existing models, datasets, and tools.

assumptions (4)
  • domain assumption An encoder-decoder architecture is particularly well suited for secure code generation because the encoder can analyze code security and the decoder can generate secure code.
    Invoked in Section 7.1 as the basis for choosing CodeT5-style models; no comparative experiments or quantitative justification are provided.
  • domain assumption Fine-tuning a base model on security-oriented datasets such as CWE, OWASP, and NVD transfers to secure generation on real-world prompts.
    Section 7.2 assumes this transfer when proposing fine-tuning with datasets from Section 5.2.2; the paper cites the datasets but does not measure transfer.
  • domain assumption Automated security analysis tools and sandbox execution can correctly label generated code, and these labels can serve as reliable training signals in a feedback loop.
    Sections 7.2 and 7.3 build the continuous training loop on this assumption. The paper itself notes in Section 4.2 that many tools cannot analyze incomplete code.
  • domain assumption Re-feeding user prompts and generated code back into training improves future output quality without introducing new privacy or security problems.
    Section 7.3 proposes recording prompts and code for continuous training; the paper mentions user consent in a footnote but provides no analysis of data-quality or privacy effects.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Optimizing AI-Assisted Code Generation." pith.science (2026). https://pith.science/paper/6FCRTB7L

@misc{pith2026241210953,
  author       = {Pith},
  title        = {Pith review of: Optimizing AI-Assisted Code Generation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6FCRTB7L}},
  note         = {Machine review of arXiv:2412.10953}
}
read the original abstract

In recent years, the rise of AI-assisted code-generation tools has significantly transformed software development. While code generators have mainly been used to support conventional software development, their use will be extended to powerful and secure AI systems. Systems capable of generating code, such as ChatGPT, OpenAI Codex, GitHub Copilot, and AlphaCode, take advantage of advances in machine learning (ML) and natural language processing (NLP) enabled by large language models (LLMs). However, it must be borne in mind that these models work probabilistically, which means that although they can generate complex code from natural language input, there is no guarantee for the functionality and security of the generated code. However, to fully exploit the considerable potential of this technology, the security, reliability, functionality, and quality of the generated code must be guaranteed. This paper examines the implementation of these goals to date and explores strategies to optimize them. In addition, we explore how these systems can be optimized to create safe, high-performance, and executable artificial intelligence (AI) models, and consider how to improve their accessibility to make AI development more inclusive and equitable.

Figures

Figures reproduced from arXiv: 2412.10953 by the authors.

Figure 1
Figure 1. The Transformer - model architecture. [92]. Another interesting development is the increasing use of platforms such as Hugging Face [19] and other LLM hubs, which are central points of contact for pre-trained models and tools for fine-tuning and integrating LLMs [84]. These platforms provide access not only to models but also to the infrastructure to adapt and use them efficiently for different tasks. This enables a… view at source ↗
Figure 2
Figure 2. This example shows how ChatGPT effectively solves a simple problem by generating a comprehensive Python script to manage a list of events. The script includes detailed comments to aid understanding and make it easier for non-experts to follow along. ChatGPT also provides a detailed explanation of how the code works and a sample output. On the economic side, the use of such AI techniques improves the efficiency and q… view at source ↗
Figure 3
Figure 3. This example shows how ChatGPT effectively solves a problem using deep learning techniques by generating a comprehensive Python script to classify the Iris dataset. The output includes details of the process and code functionality, required libraries, useful side information and the code itself. value for junior developers. In addition, the models created by non-AI experts and laypeople must be prepared in such a wa… view at source ↗
Figures from the paper (1 more)
Figure 5
Figure 5. Figure 5: Design of the trainings loop. Left side: Initial training of the base model. Middle: Zero-Trust training of the base model. Right: Continues training during life time. code. As a complement to an encoder-decoder model, they can help to analyze the code in detail and id…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Bespoke Visual Assistance: What and How do Blind and Low-Vision People Create with Agentic Programming?

    cs.HC 2026-07 conditional novelty 7.0 of 10

    Five blind and low-vision co-designers, using the agentic programming tool ProgramAT, created 37 custom camera-based assistive tools, revealing motivations, iterative strategies, and challenges like model limits and s...

Reference graph

Works this paper leans on

101 extracted references · 34 canonical work pages · cited by 1 Pith paper

  1. [86]

    Journal ofTheColloquiumforInformationSystemsSecurityEducation11,6

    Taeb,M.,Chi,H.,Bernadin,S.,2024.Assessingtheeffectivenessandsecurityimplicationsofaicodegenerators. Journal ofTheColloquiumforInformationSystemsSecurityEducation11,6. URL: https://cisse.info/journal/index. php/cisse/article/view/180, doi:10.53735/cisse.v11i1.180

  2. [1]

    Howreadableismodel-generatedcode?examiningreadabilityandvisualinspectionofgithubcopilot, in: ASE ’22, Association for Computing Machinery, New York, NY

    AlMadi,N.,2022. Howreadableismodel-generatedcode?examiningreadabilityandvisualinspectionofgithubcopilot, in: ASE ’22, Association for Computing Machinery, New York, NY. pp. 1–5. doi:10.1145/3551349.3560438

  3. [2]

    State of developer experience report 2024 | atlassian

    Atlassian, 2024-08-02. State of developer experience report 2024 | atlassian. URL:https://www.atlassian.com/ software/compass/resources/state-of-developer-2024. Torka et al.:Preprint submitted to Elsevier Page 27 of 32 Optimizing AI-Assisted Code Generation: Enhancing Security, Efficiency, and Accessibility in Software Development

  4. [3]

    Program synthesis with large language models

    Austin, J., Odena, A., Nye, M., Bosma, M., Michalewski, H., Dohan, D., Jiang, E., Cai, C., Terry, M., Le Quoc, Sutton, C., . Program synthesis with large language models. URL:http://arxiv.org/pdf/2108.07732

  5. [4]

    Deepcode ai fix: Fixing security vulnerabilities with large language models

    Berabi, B., Gronskiy, A., Raychev, V., Sivanrupan, G., Chibotaru, V., Vechev, M., . Deepcode ai fix: Fixing security vulnerabilities with large language models. URL:http://arxiv.org/pdf/2402.13291

  6. [5]

    Gpt-neox-20b: An open-source autoregressive language model

    Black, S., Biderman, S., Hallahan, E., Anthony, Q., Gao, L., Golding, L., He, H., Leahy, C., McDonell, K., Phang, J., Pieler, M., Prashanth, U.S., Purohit, S., Reynolds, L., Tow, J., Wang, B., Weinbach, S., 2022. Gpt-neox-20b: An open-source autoregressive language model. URL:https://arxiv.org/abs/2204.06745, doi:10.48550/ARXIV. 2204.06745

  7. [6]

    Kaggle forecasting competitions: An overlooked learning opportunity

    Bojer, C.S., Meldgaard, J.P., 2021. Kaggle forecasting competitions: An overlooked learning opportunity. Interna- tional Journal of Forecasting 37, 587–603. URL: https://www.sciencedirect.com/science/article/pii/ S0169207020301114, doi:10.1016/j.ijforecast.2020.07.007

  8. [7]

    Artificial intelligence takes center stage: exploring the capabilities and implications of chatgpt and other ai-assisted technologies in scientific research and education

    Borger, J.G., Ng, A.P., Anderton, H., Ashdown, G.W., Auld, M., Blewitt, M.E., Brown, D.V., Call, M.J., Collins, P., Freytag, S., Harrison, L.C., Hesping, E., Hoysted, J., Johnston, A., McInneny, A., Tang, P., Whitehead, L., Jex, A., Naik, S.H., 2023. Artificial intelligence takes center stage: exploring the capabilities and implications of chatgpt and oth...

Show all 101 references
  1. [8]

    Generative ai degrades online communities

    Burtch, G., Lee, D., Chen, Z., 2024. Generative ai degrades online communities. Communications of the ACM 67, 40–42. doi:10.1145/3624732

  2. [10]

    Evaluating large language models trained on code

    Chen, M., Tworek, J., Jun, H., Yuan, Q., Pinto, H.P.d.O., Kaplan, J., Edwards, H., Burda, Y., Joseph, N., Brockman, G., Ray, A., Puri, R., Krueger, G., Petrov, M., Khlaaf, H., Sastry, G., Mishkin, P., Chan, B., Gray, S., Ryder, N., Pavlov, M., Power, A., Kaiser, L., Bavarian, ...

  3. [11]

    Devaic: A tool for security assessment of ai-generated code

    Cotroneo, D., de Luca, R., Liguori, P., . Devaic: A tool for security assessment of ai-generated code. URL:http: //arxiv.org/pdf/2404.07548

  4. [12]

    Forecasting the m4 competition weekly data: Forecast pro’s winning approach

    Darin, S.G., Stellwagen, E., 2020. Forecasting the m4 competition weekly data: Forecast pro’s winning approach. International Journal of Forecasting 36, 135–141. doi:10.1016/j.ijforecast.2019.03.018

  5. [13]

    Focus: Google, one of ai’s biggest backers, warns own staff about chatbots

    Dastin, J., Tong, A., 2023-06-15. Focus: Google, one of ai’s biggest backers, warns own staff about chatbots. Reuters Media URL: https://www.reuters.com/technology/ google-one-ais-biggest-backers-warns-own-staff-about-chatbots-2023-06-15/

  6. [14]

    Ai assisted data labeling with interactive auto label

    Desmond, M., Brachman, M., Duesterwald, E., Dugan, C., Nath Joshi, N., Pan, Q., Spina, C., 2022. Ai assisted data labeling with interactive auto label. Proceedings of the AAAI Conference on Artificial Intelligence 36, 13161–13163. URL: https://ojs.aaai.org/index.php/AAAI/artic...

  7. [15]

    The economic impact of the ai-powered developer lifecycle and lessons from github copilot

    Dohmke, T., 2023. The economic impact of the ai-powered developer lifecycle and lessons from github copilot. URL: https://github.blog/news-insights/research/ the-economic-impact-of-the-ai-powered-developer-lifecycle-and-lessons-from-github-copilot/

  8. [16]

    Seachangeinsoftwaredevelopment:Economicandproductivityanalysisofthe ai-powered developer lifecycle

    Dohmke,T.,Iansiti,M.,Richards,G.,. Seachangeinsoftwaredevelopment:Economicandproductivityanalysisofthe ai-powered developer lifecycle. URL:http://arxiv.org/pdf/2306.15033v1

  9. [17]

    A european ai on demand platform and ecosystem

    Europäischen Kommission, 2023-11-14. A european ai on demand platform and ecosystem. URL:https://cordis. europa.eu/project/id/825619/reporting/de

  10. [18]

    Home page | ai-on-demand

    European Union, 2023. Home page | ai-on-demand. URL:https://www.ai4europe.eu/

  11. [19]

    Hugging face – the AI community building the future

    Face, H., . Hugging face – the AI community building the future. URL:https://huggingface.co/

  12. [20]

    A c/c++ code vulnerability dataset with code changes and cve summaries, in: MSR ’20: Proceedings of the 17th International Conference on Mining Software Repositories, ACM

    Fan, J., Li, Y., Wang, S., Nguyen, T.N., 2020. A c/c++ code vulnerability dataset with code changes and cve summaries, in: MSR ’20: Proceedings of the 17th International Conference on Mining Software Repositories, ACM. doi:10.1145/3379597.3387501

  13. [21]

    Towards ai assisted domain modeling, in: Reinhartz- Berger, I., Sadiq, S

    Feltus, C., Ma, Q., Proper, H.A., Kelsen, P., 2021. Towards ai assisted domain modeling, in: Reinhartz- Berger, I., Sadiq, S. (Eds.), Advances in Conceptual Modeling, Springer International Publishing, Cham. pp. 75–89. URL: https://link.springer.com/chapter/10.1007/978-3-030-8...

  14. [22]

    Feng,Z.,Guo,D.,Tang,D.,Ruan,N.,Feng,X.,Gong,M.,Shou,L.,Qin,B.,Liu,T.,Jiang,D.,Zhou,M.,2020.Codebert: A pre-trained model for programming and natural languages.arXiv:2002.08155

  15. [23]

    Fredrikson, M., Jha, S., Ristenpart, T., . Model inversion attacks that exploit confidence information and basic countermeasures, in: Proceedings of the 22nd ACM SIGSAC Conference on Computer and Communications Security, Association for Computing Machinery. pp. 1322–1333. URL:...

  16. [24]

    Incoder: A generative model for code infilling and synthesis

    Fried, D., Aghajanyan, A., Lin, J., Wang, S., Wallace, E., Shi, F., Zhong, R., Yih, W.t., Zettlemoyer, L., Lewis, M., . Incoder: A generative model for code infilling and synthesis. URL:http://arxiv.org/pdf/2204.05999v3

  17. [25]

    Them4forecastingcompetition–apractitioner’sview

    Fry,C.,Brundage,M.,2020. Them4forecastingcompetition–apractitioner’sview. InternationalJournalofForecasting 36, 156–160. doi:10.1016/j.ijforecast.2019.02.013

  18. [26]

    Gemini - chat to supercharge your ideas

    Gemini, 2024-08-06. Gemini - chat to supercharge your ideas. URL:https://gemini.google.com/app

  19. [27]

    Github - pycqa/bandit: Bandit is a tool designed to find common security issues in python code

    GitHub, 2024-08-09a. Github - pycqa/bandit: Bandit is a tool designed to find common security issues in python code. URL: https://github.com/PyCQA/bandit/tree/main

  20. [28]

    Github-python-security/pyt:Astaticanalysistoolfordetectingsecurityvulnerabilitiesinpython web applications

    GitHub,2024-08-09b. Github-python-security/pyt:Astaticanalysistoolfordetectingsecurityvulnerabilitiesinpython web applications. URL:https://github.com/python-security/pyt

  21. [29]

    Github-semgrep/semgrep:Lightweightstaticanalysisformanylanguages.findbugvariantswith patterns that look like source code

    GitHub,2024-08-09c. Github-semgrep/semgrep:Lightweightstaticanalysisformanylanguages.findbugvariantswith patterns that look like source code. URL:https://github.com/semgrep/semgrep

  22. [30]

    Github copilot· your ai pair programmer

    GitHub, 2024-09-19. Github copilot· your ai pair programmer. URL:https://github.com/features/copilot

  23. [31]

    GitHub, I., 2021. Codeql. URL:https://codeql.github.com/

  24. [32]

    The Effects of AI Assisted Programming in Software Engineering

    Gottlander, J., Khademi, T., 2023. The Effects of AI Assisted Programming in Software Engineering. URL:https: //odr.chalmers.se/items/acf7c5fa-50dc-4c11-842b-d1414e7bf689

  25. [33]

    BadNets: Identifying vulnerabilities in the machine learning model supply chain

    Gu, T., Dolan-Gavitt, B., Garg, S., . BadNets: Identifying vulnerabilities in the machine learning model supply chain. URL: http://arxiv.org/abs/1708.06733, doi:10.48550/arXiv.1708.06733, arXiv:1708.06733

  26. [34]

    GraphCodeBERT: Pre-training code representations with data flow

    Guo, D., Ren, S., Lu, S., Feng, Z., Tang, D., Liu, S., Zhou, L., Duan, N., Svyatkovskiy, A., Fu, S., Tufano, M., Deng, S.K., Clement, C., Drain, D., Sundaresan, N., Yin, J., Jiang, D., Zhou, M., . GraphCodeBERT: Pre-training code representations with data flow. URL:http://arxi...

  27. [35]

    From chatgpt to threatgpt: Impact of generative ai in cybersecurity and privacy

    Gupta, M., Akiri, C., Aryal, K., Parker, E., Praharaj, L., 2023. From chatgpt to threatgpt: Impact of generative ai in cybersecurity and privacy. IEEE Access 11, 80218–80245. doi:10.1109/ACCESS.2023.3300381

  28. [36]

    Samsung bans chatgpt, google bard, other generative ai use by staff after leak

    Gurman, M., 2023-05-02. Samsung bans chatgpt, google bard, other generative ai use by staff after leak. Bloomberg URL: https://www.bloomberg.com/news/articles/2023-05-02/ samsung-bans-chatgpt-and-other-generative-ai-use-by-staff-after-leak

  29. [38]

    How secure is ai-based coding?: A security analysis using stride and data flow diagrams, in: 2023 IEEE Virtual Conference on Communications (VCC), IEEE, Piscataway, New Jersey

    Hossain, I., Istiak Chowdhury, N.M., Hasan, R., 2023. How secure is ai-based coding?: A security analysis using stride and data flow diagrams, in: 2023 IEEE Virtual Conference on Communications (VCC), IEEE, Piscataway, New Jersey. pp. 56–61. doi:10.1109/VCC60689.2023.10474718

  30. [39]

    Improta, C., 2023. Poisoning programs by un-repairing code: Security concerns of ai-generated code, in: 2023 IEEE 34th International Symposium on Software Reliability Engineering Workshop, IEEE, Piscataway, NJ. pp. 128–131. doi:10.1109/ISSREW60843.2023.00060

  31. [40]

    Usingaiassistantsinsoftwaredevelopment:Aqualitativestudy on security practices and concerns

    Klemmer,J.H.,Horstmann,S.A.,Patnaik,N.,Ludden,C.,Burton,JR,C.,Powers,C.,Massacci,F.,Rahman,A.,Votipka, D.,Lipford,H.R.,Rashid,A.,Naiakshina,A.,Fahl,S.,. Usingaiassistantsinsoftwaredevelopment:Aqualitativestudy on security practices and concerns. URL:http://arxiv.org/pdf/2405.06371

  32. [41]

    Becoming good at ai for good, in: Fourcade, M

    Kshirsagar,M.,Robinson,C.,Yang,S.,Gholami,S.,Klyuzhin,I.,Mukherjee,S.,Nasir,M.,Ortiz,A.,Oviedo,F.,Tanner, D., Trivedi, A., Xu, Y., Zhong, M., Dilkina, B., Dodhia, R., Lavista Ferres, J.M., 2021. Becoming good at ai for good, in: Fourcade, M. (Ed.), Proceedings of the 2021 AAAI...

  33. [42]

    Coderl: Mastering code generation through pretrained models and deep reinforcement learning

    Le, H., Wang, Y., Gotmare, A.D., Savarese, S., Hoi, S.C., 2022. Coderl: Mastering code generation through pretrained models and deep reinforcement learning. arXiv preprint abs/2207.01780

  34. [43]

    AlphaCode 2 technical report

    Leblond,R.,Altché,F.,Gimeno,F.,Saade,A.,Ruddock,A.,Tallec,C.,Powell,G.,Grill,J.B.,Mikuła,M.,Lochbrunner, M., Mathieu, M., Caron, P., Shrivastava, D., Mitchell, E., Margand, G., Kelly, J., Sygnowski, J., Keeling, J., Chung, J., Kushman, N., Savinov, N., Yotov, P., Igwe, T.P., S...

  35. [44]

    Camel:Communicativeagentsfor"mind"exploration of large language model society

    Li,G.,Hammoud,H.,Itani,H.,Khizbullin,D.,Ghanem,B.,2023. Camel:Communicativeagentsfor"mind"exploration of large language model society. Advances in Neural Information Processing Systems 36, 51991–52008

  36. [45]

    Starcoder: may the source be with you! URL:http://arxiv.org/pdf/2305.06161v2

    Li, R., Allal, L.B., Zi, Y., Muennighoff, N., Kocetkov, D., Mou, C., Marone, M., Akiki, C., Li, J., Chim, J., Liu, Q., Zheltonozhskii, E., Zhuo, T.Y., Wang, T., Dehaene, O., Davaadorj, M., Lamy-Poirier, J., Monteiro, J., Shliazhko, O., Gontier, N., Meade, N., Zebaze, A., Yee, ...

  37. [46]

    Prefix-tuning: Optimizing continuous prompts for generation

    Li, X.L., Liang, P., 2021. Prefix-tuning: Optimizing continuous prompts for generation. Proceedings of the 59th AnnualMeetingoftheAssociationforComputationalLinguisticsandthe11thInternationalJointConferenceonNatural Language Processing (Volume 1: Long Papers) , 4582–4597URL:ht...

  38. [47]

    Competition-level code generation with alphacode

    Li,Y.,Choi,D.,Chung,J.,Kushman,N.,Schrittwieser,J.,Leblond,R.,Eccles,T.,Keeling,J.,Gimeno,F.,DalLago,A., Hubert, T., Choy, P., de Masson d’Autume, C., Babuschkin, I., Chen, X., Huang, P.S., Welbl, J., Gowal, S., Cherepanov, A., Molloy, J., Mankowitz, D.J., Sutherland Robson, E...

  39. [48]

    Can we generate shellcodes via natural language? an empirical study

    Liguori, P., Al-Hossami, E., Cotroneo, D., Natella, R., Cukic, B., Shaikh, S., 2022. Can we generate shellcodes via natural language? an empirical study. Automated Software Engineering 29, 1–34. URL:https://link.springer. com/article/10.1007/s10515-022-00331-3, doi:10.1007/s10...

  40. [49]

    (Ed.), 2021 IEEE 32nd International Symposium on Software Reliability Engineering, IEEE, Piscataway, NJ

    Liguori,P.,Al-Hossami,E.,Orbinato,V.,Natella,R.,Shaikh,S.,Cotroneo,D.,Cukic,B.,2021.Evil:Exploitingsoftware via natural language, in: Jin, Z. (Ed.), 2021 IEEE 32nd International Symposium on Software Reliability Engineering, IEEE, Piscataway, NJ. pp. 321–332. doi:10.1109/ISSRE...

  41. [50]

    Codexglue: A machine learning benchmark dataset for code understanding and generation

    Lu, S., Guo, D., Ren, S., Huang, J., Svyatkovskiy, A., Blanco, A., Clement, C., Drain, D., Jiang, D., Tang, D., Li, G., Zhou, L., Shou, L., Zhou, L., Tufano, M., Gong, M., Zhou, M., Duan, N., Sundaresan, N., Deng, S.K., Fu, S., Liu, S., . Codexglue: A machine learning benchmar...

  42. [51]

    On the robustness of code generation techniques: An empirical study on github copilot, in: IEEE/ACM 45th International Conference on Software Engineering (ICSE), IEEE

    Mastropaolo, A., Pascarella, L., Guglielmi, E., Ciniselli, M., Scalabrino, S., Oliveto, R., Bavota, G., 2023. On the robustness of code generation techniques: An empirical study on github copilot, in: IEEE/ACM 45th International Conference on Software Engineering (ICSE), IEEE....

  43. [52]

    Microsoft copilot: Ihr täglicher ki-begleiter

    Microsoft Copilot: Ihr täglicher KI-Begleiter, 2024-08-06. Microsoft copilot: Ihr täglicher ki-begleiter. URL:https: //copilot.microsoft.com/

  44. [53]

    Microsoft codexglue

    Microsoft Corporation, 2024. Microsoft codexglue. URL: https://github.com/microsoft/CodeXGLUE/tree/ main?tab=readme-ov-file

  45. [54]

    Ai-assisted knowledge assessment techniques for adaptive learning environments

    Minn, S., 2022. Ai-assisted knowledge assessment techniques for adaptive learning environments. Computers and Education: Artificial Intelligence 3, 100050. URL:https://www.sciencedirect.com/science/article/pii/ S2666920X22000054, doi:10.1016/j.caeai.2022.100050

  46. [55]

    Githubcopilotaipair programmer:Assetorliability? JournalofSystemsandSoftware203,111734

    MoradiDakhel,A.,Majdinasab,V.,Nikanjam,A.,Khomh,F.,Desmarais,M.C.,Jiang,Z.M.,2023. Githubcopilotaipair programmer:Assetorliability? JournalofSystemsandSoftware203,111734. URL: https://www.sciencedirect. com/science/article/pii/S0164121223001292, doi:10.1016/j.jss.2023.111734

  47. [56]

    (Eds.), Proceedings of the CHI Conference on Human Factors in Computing Systems, Association for Computing Machinery, Erscheinungsort nicht ermittelbar

    Mozannar,H.,Bansal,G.,Fourney,A.,Horvitz,E.,2024.Readingbetweenthelines:Modelinguserbehaviorandcostsin ai-assisted programming, in: Mueller, F.F., Kyburz, P., Williamson, J.R., Sas, C., Wilson, M.L., Dugas, P.T., Shklovski, I. (Eds.), Proceedings of the CHI Conference on Human...

  48. [57]

    A systematic literature review on the impact of ai models on the security of code generation

    Negri-Ribalta, C., Geraud-Stewart, R., Sergeeva, A., Lenzini, G., 2024. A systematic literature review on the impact of ai models on the security of code generation. Frontiers in Big Data 7, 1386720. URL:https://www.frontiersin. org/articles/10.3389/fdata.2024.1386720/full, do...

  49. [58]

    NeilAPerry/do-users-write-more-insecure-code-with-AI-assistants

    NeilAPerry, . NeilAPerry/do-users-write-more-insecure-code-with-AI-assistants. URL: https://github. com/NeilAPerry/Do-Users-Write-More-Insecure-Code-with-AI-Assistants . original-date: 2023-09- 06T19:00:53Z

  50. [59]

    How to jailbreak chatgpt with these prompts

    Ng, A., 2023-08-06. How to jailbreak chatgpt with these prompts. MLYearning URL:https://www.mlyearning. org/how-to-jailbreak-chatgpt/

  51. [60]

    Megavul: A c/c++ vulnerability dataset with comprehensive code representations, in: 2024 IEEE/ACM 21st International Conference on Mining Software Repositories (MSR), pp

    Ni, C., Shen, L., Yang, X., Zhu, Y., Wang, S., 2024. Megavul: A c/c++ vulnerability dataset with comprehensive code representations, in: 2024 IEEE/ACM 21st International Conference on Mining Software Repositories (MSR), pp. 738–742

  52. [61]

    Codegen: An open large language model for code with multi-turn program synthesis

    Nijkamp, E., Pang, B., Hayashi, H., Tu, L., Wang, H., Zhou, Y., Savarese, S., Xiong, C., . Codegen: An open large language model for code with multi-turn program synthesis. URL:http://arxiv.org/pdf/2203.13474v5

  53. [62]

    Crossvul: a cross-language vulnerability dataset with commit data, in: ESEC/FSE 2021, ACM

    Nikitopoulos, G., Dritsa, K., Louridas, P., Mitropoulos, D., 2021. Crossvul: a cross-language vulnerability dataset with commit data, in: ESEC/FSE 2021, ACM. doi:10.1145/3468264.3473122

  54. [63]

    CodexLeaks: Privacy leaks from code generation language models in GitHub copilot, in: Proceedings of the 32nd USENIX Security Symposium, USENIX Association, Berkeley, CA

    Niu, L., Mirza, S., Maradni, Z., Pöpper, C., 2023. CodexLeaks: Privacy leaks from code generation language models in GitHub copilot, in: Proceedings of the 32nd USENIX Security Symposium, USENIX Association, Berkeley, CA. pp. 2133–2150. URL: https://www.usenix.org/conference/u...

  55. [64]

    Openai chatgpt

    OpenAI, 2024-09-19. Openai chatgpt. URL:https://openai.com/index/chatgpt/

  56. [65]

    The convergence of ai/ml and devsecops: Revolutionizing software development

    Pakalapati, N., Venkatasubbu, S., Sistla, S.M.K., 2023. The convergence of ai/ml and devsecops: Revolutionizing software development. Journal of Knowledge Learning and Science Technology ISSN: 2959-6386 (online) 2, 189–212. URL: http://jklst.org/index.php/home/article/view/144...

  57. [66]

    The limitations of deep learning in adversarial settings, in: 2016 IEEE European Symposium on Security and Privacy (EuroS&P), pp

    Papernot, N., McDaniel, P., Jha, S., Fredrikson, M., Celik, Z.B., Swami, A., . The limitations of deep learning in adversarial settings, in: 2016 IEEE European Symposium on Security and Privacy (EuroS&P), pp. 372–387. URL: https://ieeexplore.ieee.org/abstract/document/7467366,...

  58. [67]

    Copilot CWE Scenarios Dataset

    Pearce, H., Ahmad, B., Tan, B., Dolan-Gavitt, B., Karri, R., 2021. Copilot CWE Scenarios Dataset. Zenodo. URL: https://zenodo.org/records/5225651, doi:10.5281/zenodo.5225651

  59. [68]

    Asleep at the keyboard? assessing the security of github copilot’s code contributions, in: 2022 IEEE Symposium on Security and Privacy (SP), IEEE

    Pearce, H., Ahmad, B., Tan, B., Dolan-Gavitt, B., Karri, R., 2022. Asleep at the keyboard? assessing the security of github copilot’s code contributions, in: 2022 IEEE Symposium on Security and Privacy (SP), IEEE. doi:10.1109/ sp46214.2022.9833571. Torka et al.:Preprint submit...

  60. [69]

    Examining zero-shot vulnerability repair with large language models, in: Zhao, Z

    Pearce, H., Tan, B., Ahmad, B., Karri, R., Dolan-Gavitt, B., 2023. Examining zero-shot vulnerability repair with large language models, in: Zhao, Z. (Ed.), 44th IEEE Symposium on Security and Privacy, IEEE, Piscataway, NJ. pp. 2339–

  61. [70]

    Douserswritemoreinsecurecodewithaiassistants?,in:Meng, W., Jensen, C.D., Cremers, C., Kirda, E

    Perry,N.,Srivastava,M.,Kumar,D.,Boneh,D.,2023. Douserswritemoreinsecurecodewithaiassistants?,in:Meng, W., Jensen, C.D., Cremers, C., Kirda, E. (Eds.), Proceedings of the 2023 ACM SIGSAC Conference on Computer and Communications Security, Association for Computing Machinery, Er...

  62. [71]

    Machine Teaching by Domain Experts: Towards More Humane,Inclusive, and Intelligent Machine Learning Systems

    Pinhanez, C., 2019. Machine Teaching by Domain Experts: Towards More Humane,Inclusive, and Intelligent Machine Learning Systems

  63. [72]

    Ai assisted programming, in: Choppella, V

    Rajamani, S., 2022. Ai assisted programming, in: Choppella, V. (Ed.), Proceedings of the 15th Annual ACM India Compute Conference, Association for Computing Machinery, New York,NY,United States. p. 5. doi: 10.1145/ 3561833.3568496

  64. [73]

    Chatgpt: A comprehensive review on background, applications, key challenges, bias, ethics, limitations andfuturescope

    Ray, P.P., 2023. Chatgpt: A comprehensive review on background, applications, key challenges, bias, ethics, limitations andfuturescope. InternetofThingsandCyber-PhysicalSystems3,121–154. doi: 10.1016/j.iotcps.2023.04.003

  65. [74]

    Enhancingsecurityofai-basedcodesynthesis with github copilot via cheap and efficient prompt-engineering

    Res,J.,Homoliak,I.,Perešíni,M.,Smrčka,A.,Malinka,K.,Hanacek,P.,. Enhancingsecurityofai-basedcodesynthesis with github copilot via cheap and efficient prompt-engineering. URL:http://arxiv.org/pdf/2403.12671

  66. [75]

    Lostatc:Auserstudyonthesecurityim- plicationsoflargelanguagemodelcodeassistants,in:Proceedingsofthe32ndUSENIXSecuritySymposium,USENIX Association, Berkeley, CA

    Sandoval,G.,Pearce,H.,Nys,T.,Karri,R.,Garg,S.,Dolan-Gavitt,B.,2023. Lostatc:Auserstudyonthesecurityim- plicationsoflargelanguagemodelcodeassistants,in:Proceedingsofthe32ndUSENIXSecuritySymposium,USENIX Association, Berkeley, CA. pp. 2205–2222. URL:https://www.usenix.org/confer...

  67. [76]

    What is it like to program with artificial intelligence? URL:https://www.ppig.org/papers/2022-ppig-33rd-sarkar/

    Sarkar, A., Gordon, A.D., Negreanu, C., Poelitz, C., Ragavan, S.S., Zorn, B., 2022. What is it like to program with artificial intelligence? URL:https://www.ppig.org/papers/2022-ppig-33rd-sarkar/

  68. [77]

    Integrating Structured and Unstructured Knowledge Sources for Domain-Specific Chat- bots

    Sarkar, R., 2023. Integrating Structured and Unstructured Knowledge Sources for Domain-Specific Chat- bots. URL: https://aran.library.nuigalway.ie/bitstream/handle/10379/17963/rajdeep_thesis_ final.pdf?sequence=1

  69. [78]

    Ai-based modeling: Techniques, applications and research issues towards automation, intelligent and smart systems

    Sarker, I.H., 2022. Ai-based modeling: Techniques, applications and research issues towards automation, intelligent and smart systems. SN Computer Science 3, 158. URL: https://link.springer.com/article/10.1007/ s42979-022-01043-x, doi:10.1007/s42979-022-01043-x

  70. [79]

    Adversarial attacks and defenses in large language models: Old and new threats, in: Proceedings on, PMLR

    Schwinn, L., Dobre, D., Günnemann, S., Gidel, G., . Adversarial attacks and defenses in large language models: Old and new threats, in: Proceedings on, PMLR. pp. 103–117. URL:https://proceedings.mlr.press/v239/ schwinn23a.html. ISSN: 2640-3498

  71. [80]

    Membership inference attacks against machine learning models, in: 2017IEEESymposiumonSecurityandPrivacy(SP),pp.3–18

    Shokri, R., Stronati, M., Song, C., Shmatikov, V., . Membership inference attacks against machine learning models, in: 2017IEEESymposiumonSecurityandPrivacy(SP),pp.3–18. URL: https://ieeexplore.ieee.org/abstract/ document/7958568, doi:10.1109/SP.2017.41. ISSN: 2375-1207

  72. [81]

    Securityeval dataset: Mining vulnerability examples to evaluate machine learning- based code generation techniques

    Siddiq, M.L., Santos, J.C.S., 2022. Securityeval dataset: Mining vulnerability examples to evaluate machine learning- based code generation techniques. URL:https://github.com/s2e-lab/SecurityEval

  73. [82]

    Sicherheit für ki-generierten code mit snyk | snyk

    Snyk, 2024-08-09. Sicherheit für ki-generierten code mit snyk | snyk. URL:https://snyk.io/de/solutions/ secure-ai-generated-code/

  74. [83]

    Stack overflow developer survey 2023

    Stack Overflow, 2024-08-05. Stack overflow developer survey 2023. URL:https://survey.stackoverflow.co/ 2023/

  75. [84]

    Democratizingartificialintelligence:Howno-codeaicanleveragemachinelearning operations

    Sundberg,L.,Holmström,J.,2023. Democratizingartificialintelligence:Howno-codeaicanleveragemachinelearning operations. Business Horizons 66, 777–788. URL:https://www.sciencedirect.com/science/article/pii/ S0007681323000502, doi:10.1016/j.bushor.2023.04.003

  76. [85]

    Open source security & risk analysis report (ossra) | synopsys

    Synopsys, 2024. Open source security & risk analysis report (ossra) | synopsys. URL:https://www.synopsys. com/software-integrity/resources/analyst-reports/open-source-security-risk-analysis.html# introMenu

  77. [87]

    Mitre att&ck®

    The MITRE Corporation, 2024. Mitre att&ck®. URL: https://attack.mitre.org/

  78. [88]

    Llmseceval: A dataset of natural language prompts for security evaluations

    Tony, C., Mutas, M., Díaz Ferreyra, N., Scandariato, R., 2023. Llmseceval: A dataset of natural language prompts for security evaluations. URL:https://github.com/tuhh-softsec/LLMSecEval/

  79. [89]

    Stealing machine learning models via prediction {APIs}, pp

    Tramèr, F., Zhang, F., Juels, A., Reiter, M.K., Ristenpart, T., . Stealing machine learning models via prediction {APIs}, pp. 601–618. URL: https://www.usenix.org/conference/usenixsecurity16/technical-sessions/ presentation/tramer

  80. [90]

    Natural Language Processing mit Transformern: Sprachanwendungen mit Hugging Face erstellen

    Tunstall, L., Werra, L.v., Wolf, T., Géron, A., . Natural Language Processing mit Transformern: Sprachanwendungen mit Hugging Face erstellen. Animals. 1. auflage, deutsche ausgabe ed., O’Reilly

  81. [91]

    Forecasting

    UniversityofNicosia,2024-03-29. Forecasting. URL: https://www.unic.ac.cy/iff/research/forecasting/

  82. [92]

    Attention is all you need

    Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A.N., Kaiser, Ł., Polosukhin, I., 2017. Attention is all you need. Advances in Neural Information Processing Systems 30

  83. [93]

    Apple becomes the latest company to ban chatgpt for internal use

    Vigliarolo, B., 2023-05-19. Apple becomes the latest company to ban chatgpt for internal use. The Register URL: https://www.theregister.com/2023/05/19/apple_chatgpt/. Torka et al.:Preprint submitted to Elsevier Page 31 of 32 Optimizing AI-Assisted Code Generation: Enhancing Se...

  84. [94]

    Codet5:Identifier-awareunifiedpre-trainedencoder-decodermodels for code understanding and generation, in: EMNLP, Association for Computational Linguistics

    Wang,Y.,Wang,W.,Joty,S.R.,Hoi,S.C.H.,2021. Codet5:Identifier-awareunifiedpre-trainedencoder-decodermodels for code understanding and generation, in: EMNLP, Association for Computational Linguistics. pp. 8696–8708

  85. [95]

    Building bing developer assistant

    Wei, Y., Chandrasekaran, N., Gulwani Sumit, Hamad, Y., 2015. Building bing developer assistant. Microsoft. URL: https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/bing_developer_ assistant.pdf

  86. [96]

    Naturallanguagegenerationandunderstandingofbigcode for ai-assisted programming: A review

    Wong,M.F.,Guo,S.,Hang,C.N.,Ho,S.W.,Tan,C.W.,2023. Naturallanguagegenerationandunderstandingofbigcode for ai-assisted programming: A review. Entropy 25, 888. URL:https://www.mdpi.com/1099-4300/25/6/888, doi:10.3390/e25060888

  87. [97]

    A systematic evaluation of large language models of code, in: Chaudhuri, S

    Xu, F.F., Alon, U., Neubig, G., Hellendoorn, V.J., 2022. A systematic evaluation of large language models of code, in: Chaudhuri, S. (Ed.), Proceedings of the 6th ACM SIGPLAN International Symposium on Machine Programming, Association for Computing Machinery, New York,NY,Unite...

  88. [98]

    Groundinginteractivemachinelearningtooldesigninhownon-experts actually build models, in: Proceedings of the 2018 Designing Interactive Systems Conference, ACM, New York, NY, USA

    Yang,Q.,Suh,J.,Chen,N.C.,Ramos,G.,2018. Groundinginteractivemachinelearningtooldesigninhownon-experts actually build models, in: Proceedings of the 2018 Designing Interactive Systems Conference, ACM, New York, NY, USA. pp. 573–584. doi:10.1145/3196709.3196729

  89. [99]

    Assessingthequalityofgithubcopilot’scodegeneration,in:PROMISE2022, ACM

    Yetistiren,B.,Ozsoy,I.,Tuzun,E.,2022. Assessingthequalityofgithubcopilot’scodegeneration,in:PROMISE2022, ACM. doi:10.1145/3558489.3559072

  90. [100]

    AIassistedpubliccloudmanagement

    Ylitalo,A.,. AIassistedpubliccloudmanagement. Ph.D.thesis. URL: https://www.theseus.fi/handle/10024/ 850830

  91. [101]

    Whyjohnnycan’tprompt:Hownon-aiexpertstry (andfail)todesignllmprompts,in:Proceedingsofthe2023CHIConferenceonHumanFactorsinComputingSystems, ACM, New York, NY, USA

    Zamfirescu-Pereira,J.D.,Wong,R.Y.,Hartmann,B.,Yang,Q.,2023. Whyjohnnycan’tprompt:Hownon-aiexpertstry (andfail)todesignllmprompts,in:Proceedingsofthe2023CHIConferenceonHumanFactorsinComputingSystems, ACM, New York, NY, USA. pp. 1–21. doi:10.1145/3544548.3581388

  92. [102]

    Bing developer assistant: improving developer productivity by recommending sample code, in: Zimmermann, T

    Zhang, H., Jain, A., Khandelwal, G., Kaushik, C., Ge, S., Hu, W., 2016. Bing developer assistant: improving developer productivity by recommending sample code, in: Zimmermann, T. (Ed.), Proceedings of the 2016 24th ACM SIGSOFT International Symposium on Foundations of Software...

  93. [2356]

    doi:10.1109/SP46215.2023.10179324

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.