CoChat
Sign Up
Cited by 19Social Robot Interaction and HRIAI in Service InteractionsSpeech and dialogue systems

The DOI matches its Crossref record; 222 of the 222 references it cites were found in a registry or an index.

AI apology: a critical review of apology in AI systems

Open access
Harland, Hadassah, Dazeley, Richard, Senaratne, Hashini, Vamplew, Peter, Cruz, Francisco, Nakisa, Bahareh·2025·Artificial Intelligence Review
DOI 10.1007/s10462-025-11305-8
Open this paper in CoChat — free accountOpen PDF

Abstract

Apologies are a powerful tool used in human-human interactions to provide affective support, regulate social processes, and exchange information following a trust violation. The emerging field of AI apology investigates the use of apologies by artificially intelligent systems, with recent research suggesting how this tool may provide similar value in human-machine interactions. Until recently, contributions to this area were sparse, and these works have yet to be synthesised into a cohesive body of knowledge. This article provides the first synthesis and critical analysis of the state of AI apology research, focusing on studies published between 2020 and 2023. We derive a framework of attributes to describe five core elements of apology: outcome, interaction, offence, recipient, and offender. With this framework as the basis for our critique, we show how apologies can be used to recover from misalignment in human-AI interactions, and examine trends and inconsistencies within the field. Among the observations, we outline the importance of curating a human-aligned and cross-disciplinary perspective in this research, with consideration for improved system capabilities and long-term outcomes.

Overview

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1007/s10462-025-11305-8

Cites

222 of 222 references verified0 citation errors · 0 not checkable

Statistics

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 222 references

Research works printed but not linked: 25

  1. Almuttalibi N (2016) The forms and functions of apologies with indicative references references to the letters of Keats and Byron. Int J Eng Lit 6(February):11–22
  2. Amodei D, Olah C, Steinhardt J, et al. (2016) Concrete problems in AI safety. ArXiv
  3. Aydin M (2013) Cross cultural pragmatics: a study of apology speech acts by Turkish speakers, American English speakers and advance nonnative speakers of English in Turkey. PhD thesis, Minnesota State University - Mankato
  4. Benner D, Elshan E, Schöbel S, et al. (2021) What do you mean? A review on recovery strategies to overcome conversational breakdowns of conversational agents. In: Int Conf on Inf Syst (ICIS), pp 1–17
  5. Eslami-Rasekh A, Mardani M (2010) Investigating the effects of teaching apology speech act, with a focus on intensifying strategies, on pragmatic development of EFL learners: The Iranian context. Int J Lang Soc Cult 30(1):96–103
  6. Gert B, Gert J (2020) The definition of morality. In: Zalta EN (ed) Stanford Encycl of Philos, fall, 2020th edn. Stanford University, Metaphysics Research Lab
  7. Hamacher A (2015) Believing in BERT: making good on bad robot behavior. PhD thesis, University College London
  8. Harris AM, DeChurch L (2023) “Alexa, how can I trust you again?” Trust repair in human-AI teams. PhD thesis, United States– Illinois
  9. Jelínek M, Fischer K (2023) Trust regulation in social robotics: from violation to repair. [ArXiv]
  10. Joyce R (1999) Apologizing. Public Aff Q 13(2):159–173
  11. Khavas ZR (2021) A review on trust in human-robot interaction. [ArXiv]
  12. Kreiter R (2023) Can we forgive a robotic teammate? The role of trust and trust repair strategies in human-agent teams. PhD thesis, University of Twente
  13. Lazare A (2004) On apology. Oxford University Press
  14. Lee Y, Bae JE, Kwak SS, et al. (2011) The effect of politeness strategy on human - robot collaborative interaction on malfunction of robot vacuum cleaner. In RSS workshop on HRI
  15. Li M, Lee JD (2023) Measure and manage trust in human-AI conversations. PhD thesis, University of Winconsin, Madison
  16. Lu PPY, Konishi M, Sano S, et al. (2022) What makes an apology more effective? Exploring anthropomorphism, individual differences, and emotion in human-automation trust repair. [ArXiv]
  17. Na G, Choi J, Kang H (2023) It’s Not my fault, but I’m to blame: the effect of a home robot’s attribution and approach movement on trust and emotion of users. Int J Hum-Comput Interact 10:2209977
  18. Perkins R, Khavas ZR, Robinette P (2021) Trust calibration and trust respect: a method for building team cohesion in human robot teams. [ArXiv]
  19. Rzepka C, Berger B (2018) User interaction with AI-enabled systems: a systematic review of IS research. ICIS 2018 proc
  20. Soares N, Fallenstein B (2014) Aligning superintelligence with human interests: a technical research agenda. Tech. rep, Machine Intelligence Research Institute (MIRI) technical report
  21. Stowers K (2017) The role of accounts and apologies in mitigating blame toward human and machine agents. PhD thesis, University of Central Florida
  22. Wilson L, Mechenbier MX, Melonçon L (2020) Affective investment. Acad Labor: Res Artist 4(6) · DOI
  23. Wright B, Collins E, Cameron D (2022) When robots interact with groups, where does the trust reside? [ArXiv]
  24. Yu D, Li L, Su H, et al. (2023) Assessing the potential of AI-assisted pragmatic annotation: the case of apologies. [ArXiv]
  25. Zhang X (2023) “Sorry, it was my fault”: repairing trust in human-robot interactions. Int J Hum-Comput Stud 175(103031):79

Cited by

Sorted by citations

Related

See all

Metadata

Authors
Harland, H., Dazeley, R., Senaratne, H., Vamplew, P., Cruz, F., & Nakisa, B.
Affiliations
  • Commonwealth Scientific and Industrial Research Organisation (AU)
  • Deakin University (AU)
  • Federation University (AU)
  • UNSW Sydney (AU)
Corresponding author
Harland, Hadassah
ORCID
Venue
Artificial Intelligence Review
Publisher
Springer Science and Business Media LLC
Published
2025-10-06
Volume
58
Issue
12
ISSN
  • 0269-2821 (ISSN-L)
  • 1573-7462
DOI
10.1007/s10462-025-11305-8
OpenAlex
W4414849137
Cited by
19
References
222
Citations by year
2026: 15 · 2025: 4
FWCI
21.05
Type
journal article
Language
en
Open access
Licence
Locations
  • Artificial Intelligence Review (publishedVersion, cc-by)
  • arXiv (Cornell University) (submittedVersion, cc-by)
  • Figshare (submittedVersion, cc-by)
  • arXiv (Cornell University) (cc-by)
Funders
Deakin University
Topics
  • Social Sciences › Psychology › Social Psychology › Social Robot Interaction and HRI
  • Social Sciences › Business, Management and Accounting › Marketing › AI in Service Interactions
  • Physical Sciences › Computer Science › Artificial Intelligence › Speech and dialogue systems
Keywords
apology, human-AI interaction, breach of trust, crime, emotional support, social regulation

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review