‘AI lost the prompt!’ Replacing ‘AI hallucination’ to distinguish between mere errors and irregularities

José María Ariso; Peter Bannister

doi:10.1007/s00146-025-02757-1

journal article Open Access Nov 24, 2025

‘AI lost the prompt!’ Replacing ‘AI hallucination’ to distinguish between mere errors and irregularities

José María Ariso Peter Bannister

AI & SOCIETY · Springer Science and Business Media LLC

View at Publisher Save 10.1007/s00146-025-02757-1

Abstract

Abstract
One of the principal areas of current AI research concerns what are termed “hallucinations”. Whilst hundreds of different definitions and classifications of “AI hallucination” have been published, none have yet considered the distinction between errors and irregularities in Wittgenstein’s sense. This article provides a straightforward explanation of this distinction, illustrated through examples of AI outputs drawn from various publications. We then examine the terms proposed as alternatives to “hallucination” and highlight both their strengths and weaknesses. Drawing upon this analysis, we establish criteria for proposing alternative terms that encompass both errors and irregularities in Wittgenstein’s sense. Our aim is not to definitively resolve the ongoing debate surrounding suitable replacements for “AI hallucination”, but rather to provide a comprehensive overview of the characteristics and nuances that this distinction brings to the discussion. For unlike errors, irregularities prove entirely incomprehensible to users, owing to the grammatical gap created when fundamental certainties that underpin meaningful language use are violated. Given that irregularities prove incomprehensible, as they violate the certainties underlying meaningful language use, the most trustworthy AI systems may ultimately be those that recognise their own epistemic boundaries rather than those that produce seemingly perfect outputs.

Topics

No keywords indexed for this article. Browse by subject →

References

70

[1]

Ariso JM (2015) Some variations of the certainty of one’s own death. Linguist Philos Invest 14:82–96

[2]

Ariso JM (2017) Negative certainty. Educ Philos Theory 49(1):7–16 10.1080/00131857.2016.1194739

[3]

Ariso JM (2020) Religious certainty: peculiarities and pedagogical considerations. Stud Philos Educ 39(6):657–669 10.1007/s11217-020-09735-8

[4]

Ariso JM (2022a) Is there an internal link between seeing a human and seeing one to whom moral consideration is due? In: Eriksen C, Hermann J, O’Hara N, Pleasants N (eds) Philosophical perspectives on moral certainty. Routledge, New York and London, pp 212–228 10.4324/9781003178927-12

[5]

Ariso JM (2022b) The teacher as persuader: on the application of Wittgenstein’s notion of ‘Persuasion’ in educational practice. Educ Philos Theory 54(10):1621–1630 10.1080/00131857.2021.1930529

[6]

Ariso JM (2024) On why ‘trust’ constitutes an appropriate synonym for ‘certainty’ in Wittgenstein’s sense: what pupils can learn from its staging. Stud Philos Educ 43(2):163–176 10.1007/s11217-023-09924-1

[7]

Ariso JM (2025a) Hypochondriacal doubt: how it devours itself despite its seeming consistence. J Med Philos 50(3):203–211 10.1093/jmp/jhaf009

[8]

Ariso JM (2025b) They just say so! Second language teaching and the acquisition of certainties. Educ Philos Theory 57(2):177–185 10.1080/00131857.2024.2412758

[9]

Ariso JM (2025c) What do science and historical denialists deny – if any – when addressing certainties in Wittgenstein’s sense. Open Philos 8(1):1–12, art. 20250060 10.1515/opphil-2025-0060

[10]

Comparing Physician and Artificial Intelligence Chatbot Responses to Patient Questions Posted to a Public Social Media Forum

John W. Ayers, Adam Poliak, Mark Dredze et al.

JAMA Internal Medicine 2023 10.1001/jamainternmed.2023.1838

[11]

Bannister P (forthcoming) ParadAIse L0st? Forthcoming in Higher Education Research and Development

[12]

Bannister P, Carver M (2024) ‘I don’t need professional development; I want institutional development’: legitimising marginalised epistemic capital that disrupts generative AI discourse. Prof Dev Educ 51(3):547–565

[13]

BBC (2017) Google AI defeats human Go champion. https://www.bbc.com/news/technology-40042581. Accessed 8 May 2025

[14]

Bender EM, Gebru T, McMillan-Major A, Shmitchell S (2021) On the dangers of stochastic parrots: can language models be too big? In: Proceedings of the 2021 ACM conference on fairness, accountability, and transparency, pp 610–623 10.1145/3442188.3445922

[15]

Bjerring JC, Busch J (2021) Artificial intelligence and patient-centered decision-making. Philos Technol 34(2):349–371 10.1007/s13347-019-00391-6

[16]

Blair A, Saffidine A (2019) AI surpasses humans at six-player poker. Science 365(6456):864–865 10.1126/science.aay7774

[17]

Bleuler E (1916) Lehrbuch der Psychiatrie. Verlag von Julius Springer, Berlin

[18]

Blom JD (2023) A dictionary of hallucinations. Springer, Cham 10.1007/978-3-031-25248-8

[19]

Borji A (2023) A categorical archive of ChatGPT failures. arXiv. https://doi.org/10.21203/rs.3.rs-2895792/v1 10.21203/rs.3.rs-2895792/v1

[20]

Brender TD (2023) Chatbot confabulations are not hallucinations—reply. JAMA Intern Med 183(10):1177–1178 10.1001/jamainternmed.2023.3875

[21]

Bryant A (2023) AI chatbots: threat or opportunity? Informatics 10(2):49 10.3390/informatics10020049

[22]

Cambridge University Press & Assessment (n.d.) Irregularity. In: Cambridge English dictionary. https://dictionary.cambridge.org/es/diccionario/ingles/irregularity. Accessed 8 May 2025

[23]

Chanda SS, Banerjee DN (2024) Omission and commission errors underlying AI failures. AI & Soc 39:937–960 10.1007/s00146-022-01585-x

[24]

Cheng M, Blodgett SL, DeVrio A, Egede L, Olteanu A (2025) Dehumanizing machines: mitigating anthropomorphic behaviors in text generation systems. arXiv preprint. https://arxiv.org/abs/2502.14019 10.18653/v1/2025.acl-long.1259

[25]

Coliva A (2010) Moore and Wittgenstein. Scepticism, certainty and common sense. Palgrave Macmillan, Hampshire

[26]

Descartes R (1986) Meditations on first philosophy: with selections from the objections and replies. Cambridge University Press, Cambridge

[27]

Dunning B (2020) Australia doesn’t exist, and other geographic conspiracy theories. Skeptoid Podcast #745. https://skeptoid.com/episodes/4745. Accessed 17 Apr 2025

[28]

Edwards B (2023) Why ChatGPT and bing chat are so good at making things up. Ars Tecnica. https://arstechnica.com/information-technology/2023/04/why-ai-chatbots-are-the-ultimate-bs-machines-and-how-people-hope-to-fix-them/ Accessed 22 Apr 2025

[29]

ChatGPT: these are not hallucinations – they’re fabrications and falsifications

Robin Emsley

Schizophrenia 2023 10.1038/s41537-023-00379-4

[30]

Faulkner P (2007) On telling and trusting. Mind 116(464):875–902 10.1093/mind/fzm875

[31]

Frankfurt H (2005) On bullshit. Princeton University Press, Princeton 10.1515/9781400826537

[32]

Goodfellow IJ, Shlens J, Szegedy C (2015) Explaining and harnessing adversarial examples. In: ICLR 2015. https://doi.org/10.48550/arXiv.1412.6572. Accessed 8 May 2025 10.48550/arxiv.1412.6572

[33]

Grace K, Stein-Perlman Z, Weinstein-Raun B, Salvatier J (2022) 2022 expert survey on progress in AI. AI Impacts. https://aiimpacts.org/2022-expert-survey-on-progress-in-ai/. Accessed 8 May 2025

[34]

Grace K, Stewart H, Sandkühler JF, Thomas S, Weinstein-Raun B, Brauner J (2024) Thousands of AI Authors on the future of AI. AI Impacts. https://aiimpacts.org/wp-content/uploads/2023/04/Thousands_of_AI_authors_on_the_future_of_AI.pdf. Accessed 8 May 2025 10.1613/jair.1.19087

[35]

Grote T, Berens P (2020) On the ethics of algorithmic decision-making in healthcare. J Med Ethics 46(3):205–211 10.1136/medethics-2019-105586

[36]

Hatem R, Simmons B, Thornton JE (2023) A call to address AI ‘Hallucinations’ and how healthcare professionals can mitigate their risks. Cureus 15(9):art. e44720

[37]

Hawley K (2014) Trust, distrust and commitment. Nous 48(1):1–20 10.1111/nous.12000

[38]

ChatGPT is bullshit

Michael Townsen Hicks, James Humphries, Joe Slater

Ethics and Information Technology 2024 10.1007/s10676-024-09775-5

[39]

Humphreys P (2020) Predictive failures in neural nets. lecture series in evidence, model and explanations. Philosophy of Science India. https://www.youtube.com/watch?v=2VFPXbrCqzM. Accessed 8 May 2025

[40]

Izadi S, Forouzanfar M (2024) Error Correction and adaptation in conversational AI: a review of techniques and applications in chatbots. AI 5:803–841 10.3390/ai5020041

[41]

Survey of Hallucination in Natural Language Generation

Ziwei Ji, Nayeon Lee, Rita Frieske et al.

ACM Computing Surveys 10.1145/3571730

[42]

Jones K (1996) Trust as an affective attitude. Ethics 107(1):4–25 10.1086/233694

[43]

Lacker K (2020) Giving GPT-3 a Turing test. Kevin Lacker’s blog, 6 July 2020. https://lacker.io/ai/2020/07/06/giving-gpt-3-a-turing-test.html. Accessed 17 Apr 2025

[44]

Leaver T, Srdarov S (2023) ChatGPT isn’t magic: the hype and hypocrisy of generative artificial intelligence (AI) rhetoric. M/C J. https://doi.org/10.5204/mcj.3004 10.5204/mcj.3004

[45]

Lee M (2023) A mathematical investigation of hallucination and creativity in GPT models. Mathematics 11(10):2320 10.3390/math11102320

[46]

Li Z (2023) The dark side of ChatGPT: legal and ethical challenges from stochastic parrots and hallucination. arXiv preprint. https://arxiv.org/abs/2304.14347. Accessed 8 May 2025

[47]

Liu H, Xue W, Chen Y, Chen D, Zhao X, Wang K, Hou L, Li R, Peng W (2025) A survey on hallucination in large vision-language models. ACM Trans Inf Syst 43(2):1–55 10.1145/3704999

[48]

Loos E, Gröpler J, Goudeau M-LS (2023) Using ChatGPT in education: human reflection on ChatGPT’s self-reflection. Societies 13:1–18, art. 196 10.3390/soc13080196

[49]

Madden MG, McNicholas BA, Laffey JG (2023) Assessing the usefulness of a large language model to query and summarize unstructured medical notes in intensive care. Intensive Care Med 49(8):1018–1020 10.1007/s00134-023-07128-2

[50]

Maleki N, Padmanabhan B, Dutta K (2024) AI hallucinations: a misnomer worth clarifying. In: 2024 IEEE conference on artificial intelligence (CAI), pp 133–138 10.1109/cai59869.2024.00033

Showing 50 of 70 references

Metrics

2

Citations

70

References

Details

Published: Nov 24, 2025
License: View

Authors

Cite This Article

José María Ariso, Peter Bannister (2025). ‘AI lost the prompt!’ Replacing ‘AI hallucination’ to distinguish between mere errors and irregularities. AI & SOCIETY. https://doi.org/10.1007/s00146-025-02757-1

‘AI lost the prompt!’ Replacing ‘AI hallucination’ to distinguish between mere errors and irregularities

You May Also Like