pith. sign in

arxiv: 2401.15284 · v6 · submitted 2024-01-27 · 💻 cs.CY · cs.AI

Beyond principlism: Practical strategies for ethical AI use in research practices

Zhicheng Lin This is my paper

Pith reviewed 2026-05-24 04:38 UTC · model grok-4.3

classification 💻 cs.CY cs.AI
keywords ethical AIAI in researchgenerative AIpractical ethicsresearch integritytransparencybias mitigationdocumentation guidelines
0
0 comments X

The pith

Researchers need five concrete goals with strategies rather than abstract ethical principles to use AI responsibly in daily work.

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

The paper identifies a gap where existing ethical approaches for AI in research—relying on high-level principles, rigid rules, or tech fixes—offer little help for everyday decisions. It proposes a user-centered, realism-inspired method that defines five specific goals to make ethical use practical. These goals address understanding models and biases, respecting privacy and rights, avoiding violations, comparing benefits to alternatives, and ensuring transparency. Actionable strategies, misuse examples, and documentation guidelines accompany each goal. If effective, this would let researchers apply AI to speed progress while maintaining integrity through clear daily practices.

Core claim

The paper claims that ethical AI application in research requires a user-centered, realism-inspired approach that specifies five goals—understanding model training and outputs including bias mitigation; respecting privacy, confidentiality, and copyright; avoiding plagiarism and policy violations; applying AI beneficially compared to alternatives; and using AI transparently and reproducibly—each with actionable strategies, realistic misuse cases, and corrective measures, along with new documentation guidelines to improve transparency and reproducibility, rather than depending on abstract principles alone.

What carries the argument

The five goals for ethical AI use, each paired with strategies and misuse cases, inside a user-centered framework that evaluates utility against existing alternatives.

If this is right

  • Researchers gain specific steps for bias mitigation and privacy protection when incorporating AI tools.
  • Decisions on AI tools should compare their utility to non-AI methods rather than relying on isolated performance numbers.
  • Standard documentation practices will make AI contributions to research more reproducible and verifiable.
  • Training programs can adopt the five goals to build daily habits for responsible AI use.
  • Guidelines should be updated as AI capabilities change to keep supporting innovation alongside integrity.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Field-specific trials of the five goals could identify which strategies work best in areas like biology or social sciences.
  • The focus on alternatives might lead to more hybrid workflows that combine AI with traditional methods across disciplines.
  • Widespread use could lower rates of retractions tied to undisclosed or misused AI assistance.
  • The goal-list format might serve as a model for ethics guidance on other new research technologies.

Load-bearing premise

That spelling out these five goals with strategies and documentation rules will give researchers effective practical guidance where abstract principles have fallen short.

What would settle it

An audit or survey of published AI-assisted papers showing no measurable increase in transparency, reproducibility, or reduction in ethical violations among researchers who follow the five goals versus those who do not.

read the original abstract

The rapid adoption of generative artificial intelligence (AI) in scientific research, particularly large language models (LLMs), has outpaced the development of ethical guidelines, leading to a "Triple-Too" problem: too many high-level ethical initiatives, too abstract principles lacking contextual and practical relevance, and too much focus on restrictions and risks over benefits and utilities. Existing approaches--principlism (reliance on abstract ethical principles), formalism (rigid application of rules), and technological solutionism (overemphasis on technological fixes)--offer little practical guidance for addressing ethical challenges of AI in scientific research practices. To bridge the gap between abstract principles and day-to-day research practices, a user-centered, realism-inspired approach is proposed here. It outlines five specific goals for ethical AI use: 1) understanding model training and output, including bias mitigation strategies; 2) respecting privacy, confidentiality, and copyright; 3) avoiding plagiarism and policy violations; 4) applying AI beneficially compared to alternatives; and 5) using AI transparently and reproducibly. Each goal is accompanied by actionable strategies and realistic cases of misuse and corrective measures. I argue that ethical AI application requires evaluating its utility against existing alternatives rather than isolated performance metrics. Additionally, I propose documentation guidelines to enhance transparency and reproducibility in AI-assisted research. Moving forward, we need targeted professional development, training programs, and balanced enforcement mechanisms to promote responsible AI use while fostering innovation. By refining these ethical guidelines and adapting them to emerging AI capabilities, we can accelerate scientific progress without compromising research integrity.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

0 major / 3 minor

Summary. The paper identifies a 'Triple-Too' problem in ethical AI guidelines for scientific research and argues that principlism, formalism, and technological solutionism provide insufficient practical guidance. It proposes a user-centered, realism-inspired alternative consisting of five specific goals (understanding model training/output and bias mitigation; respecting privacy/confidentiality/copyright; avoiding plagiarism/policy violations; applying AI beneficially vs. alternatives; using AI transparently/reproducibly), each paired with actionable strategies, misuse cases, and corrective measures, plus new documentation guidelines for transparency.

Significance. If adopted, the enumerated goals and strategies could supply researchers with concrete, context-sensitive tools that shift emphasis from abstract restrictions to utility evaluation and reproducibility. The inclusion of realistic misuse cases is a constructive element that distinguishes the proposal from purely declarative frameworks.

minor comments (3)
  1. [Abstract] The phrase 'realism-inspired' is introduced in the abstract but would benefit from a brief explicit definition or reference to the philosophical or methodological tradition invoked, to prevent reader ambiguity.
  2. The five goals are listed clearly in the abstract; ensure the body text maintains identical numbering and wording when expanding each goal to facilitate cross-reference.
  3. The documentation guidelines section would be strengthened by including a short example template or checklist that researchers could directly adapt.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the positive assessment and recommendation to accept the manuscript. The report accurately captures the core contribution of identifying the 'Triple-Too' problem and proposing a user-centered, realism-inspired framework with five concrete goals, actionable strategies, misuse cases, and documentation guidelines. We are pleased that the emphasis on utility evaluation, reproducibility, and realistic misuse cases was recognized as distinguishing the work from purely declarative approaches.

Circularity Check

0 steps flagged

No significant circularity

full rationale

The manuscript is a normative position paper that proposes five practical goals for ethical AI use in research, accompanied by strategies and documentation guidelines. It contains no mathematical derivations, fitted parameters, equations, or predictive claims that could reduce to inputs by construction. The central argument is an advocacy position contrasting the proposal against principlism, formalism, and solutionism; this rests on explicit reasoning rather than self-citation chains or self-definitional loops. No load-bearing step invokes prior work by the same author as an unverified uniqueness theorem, and the text is self-contained against external benchmarks as a standalone proposal without requiring empirical validation within its own logic.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

No free parameters, axioms, or invented entities because the paper is a normative position piece without quantitative models, empirical data, or formal derivations.

pith-pipeline@v0.9.0 · 5809 in / 1129 out tokens · 19520 ms · 2026-05-24T04:38:17.444220+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 1 Pith paper

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

  1. Techniques for supercharging academic writing with generative AI

    cs.CY 2023-10 unverdicted novelty 2.0

    The paper outlines a conceptual framework and prompting techniques for incorporating generative AI into academic writing routines such as outlining, drafting, and editing.

Reference graph

Works this paper leans on

50 extracted references · 50 canonical work pages · cited by 1 Pith paper · 1 internal anchor

  1. [1]

    Royal Society Open Science, 2023

    Lin, Z., Why and how to embrace AI such as ChatGPT in your academic life. Royal Society Open Science, 2023. 10: p. 230658

    work page 2023

  2. [2]

    Nature, 2023

    Wang, H., et al., Scientific discovery in the age of artificial intelligence. Nature, 2023. 620(7972): p. 47-60

    work page 2023

  3. [3]

    NEJM AI, 2024

    Liang, W., et al., Can large language models provide useful feedback on research papers? A large-scale empirical analysis. NEJM AI, 2024. 1(8)

    work page 2024

  4. [4]

    AI & Society, 2023

    Fecher, B., et al., Friend or foe? Exploring the implications of large language models on the science system. AI & Society, 2023

    work page 2023

  5. [5]

    Nature Reviews Physics,

    Birhane, A., et al., Science in the age of large language models. Nature Reviews Physics,

    work page

  6. [6]

    Journal of Medical Ethics, 2024

    Jecker, N.S., et al., AI and the falling sky: Interrogating X-Risk. Journal of Medical Ethics, 2024

    work page 2024

  7. [7]

    Trends in Cognitive Sciences, 2024

    Lin, Z., Towards an AI policy framework in scholarly publishing. Trends in Cognitive Sciences, 2024. 82(2): p. 85-88

    work page 2024

  8. [8]

    AI and Ethics, 2023

    Munn, L., The uselessness of AI ethics. AI and Ethics, 2023. 3(3): p. 869-877

    work page 2023

  9. [9]

    Lee, and L.J

    Badal, K., C.M. Lee, and L.J. Esserman, Guiding principles for the responsible development of artificial intelligence tools for healthcare. Communications Medicine,

    work page

  10. [10]

    Patterns, 2023

    Correa, N.K., et al., Worldwide AI ethics: A review of 200 guidelines and recommendations for AI governance. Patterns, 2023. 4(10): p. 100857

    work page 2023

  11. [11]

    Ienca, and E

    Jobin, A., M. Ienca, and E. Vayena, The global landscape of AI ethics guidelines. Nature Machine Intelligence, 2019. 1(9): p. 389-399

    work page 2019

  12. [12]

    Ghoz, L. and M. Hendawy, An Inventory of AI ethics: Analyzing 100 documents. MSA Engineering Journal, 2023. 2(2): p. 647-675

    work page 2023

  13. [13]

    Nature Machine Intelligence, 2019

    Mittelstadt, B., Principles alone cannot guarantee ethical AI. Nature Machine Intelligence, 2019. 1(11): p. 501-507

    work page 2019

  14. [14]

    Rees, C. and B. Muller, All that glitters is not gold: Trustworthy and ethical AI principles. AI and Ethics, 2022: p. 1-14

    work page 2022

  15. [15]

    Nature Biomedical Engineering, 2023

    Chen, R.J., et al., Algorithmic fairness in artificial intelligence for medicine and healthcare. Nature Biomedical Engineering, 2023. 7(6): p. 719-742

    work page 2023

  16. [16]

    AI and Ethics,

    Prem, E., From ethical AI frameworks to tools: A review of approaches. AI and Ethics,

    work page

  17. [17]

    AI and Ethics, 2023

    Leslie, D., Does the sun rise for ChatGPT? Scientific discovery in the age of generative AI. AI and Ethics, 2023

    work page 2023

  18. [18]

    Resnik, D.B. and M. Hosseini, The ethics of using artificial intelligence in scientific research: New guidance needed for a new tool. AI and Ethics, 2024

    work page 2024

  19. [19]

    Nature Human Behaviour, 2024

    Lin, Z., How to write effective prompts for large language models. Nature Human Behaviour, 2024. 8(4): p. 611-615. ETHICAL AI IN SCIENCE 19

    work page 2024

  20. [20]

    Nature Biomedical Engineering, 2024

    Lin, Z., Techniques for supercharging academic writing with generative AI. Nature Biomedical Engineering, 2024

    work page 2024

  21. [21]

    Nature Ecology & Evolution,

    Merow, C., et al., AI chatbots can boost scientific coding. Nature Ecology & Evolution,

    work page

  22. [22]

    Nature, 2023

    Perkel, J.M., Six tips for better coding with ChatGPT. Nature, 2023. 618(7964): p. 422- 423

    work page 2023

  23. [23]

    Proceedings of the National Academy of Sciences of the United States of America, 2024

    Rathje, S., et al., GPT is an effective tool for multilingual psychological text analysis. Proceedings of the National Academy of Sciences of the United States of America, 2024. 121(34): p. e2308950121

    work page 2024

  24. [24]

    Green, B. and S. Viljoen, Algorithmic realism: Expanding the boundaries of algorithmic thought, in Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency. 2020, Association for Computing Machinery: Barcelona, Spain. p. 19–31

    work page 2020

  25. [25]

    Pham, B.-C. and S.R. Davies, What problems is the AI act solving? Technological solutionism, fundamental rights, and trustworthiness in European AI policy. Critical Policy Studies, 2024: p. 1-19

    work page 2024

  26. [26]

    Taddeo, M. and L. Floridi, How AI can be a force for good. Science, 2018. 361(6404): p. 751-752

    work page 2018

  27. [27]

    The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery

    Lu, C., et al., The AI scientist: Towards fully automated open-ended scientific discovery. arXiv:2408.06292, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  28. [28]

    Science and Engineering Ethics, 2020

    Coeckelbergh, M., Artificial Intelligence, Responsibility Attribution, and a Relational Justification of Explainability. Science and Engineering Ethics, 2020. 26(4): p. 2051- 2068

    work page 2020

  29. [29]

    Ethics and Information Technology, 2004

    Matthias, A., The responsibility gap: Ascribing responsibility for the actions of learning automata. Ethics and Information Technology, 2004. 6(3): p. 175-183

    work page 2004

  30. [30]

    Philosophy & Technology, 2022

    Seger, E., In defence of principlism in AI ethics and governance. Philosophy & Technology, 2022. 35(2): p. 45

    work page 2022

  31. [31]

    AI and Ethics, 2023

    Watkins, R., Guidance for researchers and peer-reviewers on the ethical use of Large Language Models (LLMs) in scientific research workflows. AI and Ethics, 2023

    work page 2023

  32. [32]

    Wood, and P

    Barman, K.G., N. Wood, and P. Pawlowski, Beyond transparency and explainability: On the need for adequate and contextualized user guidelines for LLM use. Ethics and Information Technology, 2024. 26(3): p. 47

    work page 2024

  33. [33]

    European Journal of Clinical Investigation, 2024

    Lin, Z., Progress and challenges in the symbiosis of AI with science and medicine. European Journal of Clinical Investigation, 2024. n/a(n/a): p. e14222

    work page 2024

  34. [34]

    Lewis, P., et al., Retrieval-augmented generation for knowledge-intensive NLP tasks, in Advances in Neural Information Processing Systems. 2020. p. 9459-9474

    work page 2020

  35. [35]

    British Journal of Educational Technology, 2024

    Lee, J., et al., The life cycle of large language models in education: A framework for understanding sources of bias. British Journal of Educational Technology, 2024. 55(5): p. 1982-2002. ETHICAL AI IN SCIENCE 20

    work page 2024

  36. [36]

    Computational Linguistics, 2024: p

    Gallegos, I.O., et al., Bias and fairness in large language models: A survey. Computational Linguistics, 2024: p. 1-83

    work page 2024

  37. [37]

    Scientific Reports, 2024

    Fang, X., et al., Bias of AI-generated content: an examination of news produced by large language models. Scientific Reports, 2024. 14(1): p. 5224

    work page 2024

  38. [38]

    arXiv:2402.02680, 2024

    Manvi, R., et al., Large language models are geographically biased. arXiv:2402.02680, 2024

    work page arXiv 2024

  39. [39]

    Li, and Z

    Chen, Z., W. Li, and Z. Lin, Characterizing patients who may benefit from mature medical AI models. PsyArXiv preprint, 2024

    work page 2024

  40. [40]

    Smith, and A

    Palmer, A., N.A. Smith, and A. Spirling, Using proprietary language models in academic research requires explicit justification. Nature Computational Science, 2023

    work page 2023

  41. [41]

    Garrett, B.L. and C. Rudin, Interpretable algorithmic forensics. Proceedings of the National Academy of Sciences of the United States of America, 2023. 120(41): p. e2301842120

    work page 2023

  42. [42]

    Ehsan, U. and M.O. Riedl, Social construction of XAI: Do we need one definition to rule them all? Patterns, 2024. 5(2): p. 100926

    work page 2024

  43. [43]

    Nature Medicine, 2023

    Banerji, C.R.S., et al., Clinical AI tools must convey predictive uncertainty for each individual patient. Nature Medicine, 2023. 29(12): p. 2996-2998

    work page 2023

  44. [44]

    NEJM AI, 2024

    Ong, J.C.L., et al., Medical ethics of large language models in medicine. NEJM AI, 2024. 1(7): p. AIra2400038

    work page 2024

  45. [45]

    Nature, 2023

    Stokel-Walker, C., ChatGPT listed as author on research papers: many scientists disapprove. Nature, 2023. 613(7945): p. 620-621

    work page 2023

  46. [46]

    Ugar, E.T. and N. Malele, Designing AI for mental health diagnosis: challenges from sub-Saharan African value-laden judgements on mental health disorders. Journal of Medical Ethics, 2024

    work page 2024

  47. [47]

    Nature Machine Intelligence, 2023

    Sharma, A., et al., Human–AI collaboration enables more empathic conversations in text- based peer-to-peer mental health support. Nature Machine Intelligence, 2023. 5(1): p. 46-57

    work page 2023

  48. [48]

    Verma, S. and J. Rubin, Fairness definitions explained, in Proceedings of the International Workshop on Software Fairness. 2018, Association for Computing Machinery: Gothenburg, Sweden. p. 1–7

    work page 2018

  49. [49]

    Science Advances, 2023

    Lira, B., et al., Using artificial intelligence to assess personal qualities in college admissions. Science Advances, 2023. 9(41): p. eadg9405

    work page 2023

  50. [50]

    npj Digital Medicine, 2024

    Goetz, L., et al., Generalization—a key challenge for responsible AI in patient-facing clinical applications. npj Digital Medicine, 2024. 7(1): p. 126

    work page 2024