AI for the Common Good?! Pitfalls, challenges, and ethics pen-testing
Open accessThe DOI matches its Crossref record; 51 of the 52 references it cites were found in a registry or an index; 10 of the 56 sources its claims cite were checked for support.
Abstract
Recently, many AI researchers and practitioners have embarked on research visions that involve doing AI for “Good”. One frequent theme in current ethical guidelines is the requirement that AI be good for all, or contribute to the common good. The paper proposes ethics pen-testing as a method for making impediments visible and attenuating them.
Overview
CoChat Summary
## TL;DR
The paper argues that “AI for the Common Good” is impeded by four recurring issues: how problems are framed, who defines them, how knowledge is treated, and how side effects and dynamics are handled. It proposes “ethics pen-testing” as an adversarial design method to surface these weaknesses and improve AI systems.
## Key findings
- The author’s exploratory survey covered 99 contributions across four venues: 24 from Data Science for Social Good 2017, 4 from SoGood 2017, 15 from the 2017 AAAI Spring Symposium on AI for Social Good, and 56 from the 2nd AI for Good Global Summit.
- In the surveyed material, the vast majority of contributions used one social goal and one computational goal; there was generally no attempt at alternative framings of the social problem, and the difference between problem versions (Q1') was not the topic of any paper.
- For stakeholder involvement, the large majority of contributions took the problem definition from a non-academic project partner; only one contribution described an explicit multistakeholder process, and another mentioned such a process followed by partners.
- On dynamics, only one contribution in the sample included an explicit consideration of possible future dynamics; the author says the other Q1–Q4 considerations are not an integral part of current operational research practices in AI/DS for Social Good.
- The paper reports that in the first three venues, 8 out of 43 contributions covered SDG #3 (Good Health and Well-being); 5 could be linked to SDG #11 (Sustainable Cities and Communities), and 2 to transportation. In the AI for Good Global Summit, 11 contributions covered SDG #3, and one track had 13 contributions on Smart Cities and another had 6 on AI and Satellite Imagery.
## Methods
- Conceptual/normative argument: the paper draws on AI ethics codes, political philosophy, and computer-science ideas to derive four lead questions about Common Good-oriented AI.
- Exploratory document survey: it reads and assesses all hyperlinked contributions from four recent AI/DS-for-good venues, coding whether they explicitly addressed the four questions.
- The survey examined content and whether contributions contained explicit information relating to Q1–Q4; the author notes the coding was exploratory and simple.
## Limitations
- The paper says the conference survey was exploratory, the coding exercise was simple, and future work will use a codebook and more coders.
- The author notes that the findings may be an artifact of publication conventions that favor unambiguity and a linear research story, and that many surveyed documents were very short; this may have hidden alternatives and process details.
- As an evident limitation from the text: the survey covers only selected conference materials from four venues, so it is not a comprehensive study of AI for the Common Good.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
8 Valid · 9 Not valid · 31 Not checked · 90 Own results and interpretations
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“and it will commit countermeasures against such biased decisions to one formalization of 'fairness', which may mean that other notions of fairness are violated [40, 41].”
What the sources say
- Supportedcites [40] [40] A. Chouldechova, Fair prediction with disparate impact: A study of bias in recidivism prediction instruments, Big Data, 2017, 5(2), 153-16310.1089/big.2016.004728632438 source
Section 2.1 (p. 2) states that when recidivism prevalence differs between groups, a test-fair score cannot also have equal false-positive and false-negative rates. The Discussion (p. 4) adds that balancing error rates will generally produce assessments that are not free from predictive bias. These passages support the claimed trade-off between formalizations of fairness. (LLM, conf=0.92)
- Not matchedcites [41] [41] J. Kleinberg, S. Mullainathan, M. Raghavan, Inherent trade-offs in the fair determination of risk scores, In: Proceedings of the 8th Conference on Innovations in Theoretical Computer Science (ITCS), 201710.1145/3219617.3219634
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
hedgedstates a possible cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The cited sources that could be checked couldn’t confirm or dispute this claim.
In the paper
“This can lead to police deciding to patrol where there have been many drugrelated arrests in the past, rather than where there has been much drug usage”
What the sources say
- Could not verifycites [40] [40] A. Chouldechova, Fair prediction with disparate impact: A study of bias in recidivism prediction instruments, Big Data, 2017, 5(2), 153-16310.1089/big.2016.004728632438 sourceExperimental — accuracy not yet measured
The Introduction (p. 1) and Assessing impact (p. 3) discuss recidivism risk assessments used in decisions about defendants, including bail, parole, and sentencing. Searches for drug-related arrests, drug usage, predictive policing, and patrol decisions found no passage addressing the claimed choice of patrol locations. (LLM, conf=0.76)
- Not matchedcites [41] [41] J. Kleinberg, S. Mullainathan, M. Raghavan, Inherent trade-offs in the fair determination of risk scores, In: Proceedings of the 8th Conference on Innovations in Theoretical Computer Science (ITCS), 201710.1145/3219617.3219634
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Could not verify
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“as shown by an exploratory study of 99 contributions to recent conferences in the field.”
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Rests on an assumption not shown here.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“However, the call to develop and deploy AI for the Common Good has, to the best of my knowledge, not yet led to research programs or publications under that name.”
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The large majority of contributions took their problem definition from the non-academic project partner, which was usually also the data provider (for example, a city council, a health agency, or a transportation agency).”
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
28 more claims
Create a free account to see all 138 claims and to verify the 16 references still open.
- It may state two separate claims as one. 41
- Probably not a substantial claim. 93
- The quote only partly states this claim. 1
Citations
Cites
boyd, D., & Crawford, K. (2012) · Information, Communication & Society · doi:10.1080/1369118x.2012.678878 · Cited by 5,479 · cited 1×
“225], see [2] for the use of provocations to encourage reflection on big data.”
BackgroundThe Open Society and Its Enemies
Popper, K. (2012) · Routledge eBooks · doi:10.4324/9780203439913 · Cited by 3,505
Exposure to ideologically diverse news and opinion on Facebook
Bakshy, E., Messing, S., & Adamic, L. A. (2015) · Science · doi:10.1126/science.aaa1160 · Cited by 3,245 · cited 1×
“Social media have recently been described as intensifying the echo-chamber effects that media often have anyway, e.g. [46, 47, 48].”
BackgroundBarocas, S., & Selbst, A. D. (2014) · doi:10.2139/ssrn.2477899 · Cited by 3,048
Private traits and attributes are predictable from digital records of human behavior
Kosinski, M., Stillwell, D., & Graepel, T. (2013) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1218772110 · Cited by 2,829 · cited 1×
“This is a lesson learned from the history of the Cambridge Analytica case, in which ideas from an academic project in which social media were mined to predict personality [60] were later supposedly used for psychometrically micro-targeted election advertising [61].”
MethodFair prediction with disparate impact: a study of bias in recidivism prediction instruments
Chouldechova, A. (2017) · Big Data · doi:10.1089/big.2016.0047 · Cited by 2,159
The age of responsibilization: on market-embedded morality
Shamir, R. (2008) · Economy and Society · doi:10.1080/03085140701760833 · Cited by 1,025
Inherent trade-offs in the fair determination of risk scores
Kleinberg, J., Mullainathan, S., & Raghavan, M. (2017) · Schloss Dagstuhl – Leibniz-Zentrum für Informatik · doi:10.4230/lipics.itcs.2017.43 · Cited by 943
Inherent Trade-Offs in the Fair Determination of Risk Scores
Kleinberg, J., Mullainathan, S., & Raghavan, M. (2016) · arXiv · doi:10.48550/arxiv.1609.05807 · Cited by 941
To save everything, click here: the folly of technological solutionism
(2013) · Choice Reviews Online · doi:10.5860/choice.51-0324 · Cited by 879
Research works printed but not linked: 18
- [3] ACM Code of Ethics and Professional Conduct, 1992, https://www.acm.org/about-acm/acm-code-of-ethics-andprofessional-conduct
- [5] The IEEE Global Initiative for Ethical Considerations in Artificial Intelligence and Autonomous Systems, Ethically Aligned Design: A Vision for Prioritizing Human Wellbeing with Artificial Intelligence and Autonomous Systems, Version 1 for public discussion, 2016, http://standards.ieee.org/develop/indconn/ec/ead_v1.pdf
- [16] B. Mols, AI for the Common Good, ACM News, Jun. 20, 2017, https://cacm.acm.org/news/218696-ai-for-the-commongood/fulltext
- [17] A. Tanweer, B. Fiore-Gartland, Cross-sector collaboration in Data Science for Social Good: Opportunities, challenges, and open questions raised by working with academic researchers, In: Data Science for Social Good Conference, Sep. 28-29, 2017, Chicago, IL, http://dssg.uchicago.edu/wp-content/uploads/2017/09/tanweer.pdf
- [22] S. J. Russell, P. Norvig, Artificial Intelligence: A Modern Approach, Prentice Hall, Englewood Cliffs, NJ, 1995
- [49] T. T. Nguyen, P.-M. Hui, F. M. Harper, L. Terveen, J. A. Konstan, J. A., Exploring the filter bubble: The effect of using recommender systems on content diversity, In: Proceedings of International World Wide Web Conference Committee (WWW’14), ACM, New York, 201410.1145/2566486.2568012 · DOI · Paper page
- [57] G. Rockwell, B. Berendt, Information wants to be free, or does it? The ethics of datafication, Electronic Book Review, 2017, http://electronicbookreview.com/thread/technocapitalism/datafiction
- [67] R. Gavaldà, I. Koprinska, S. Kramer (Eds.), Proceedings of the Second Workshop on Data Science for Social Good co-located with European Conference on Machine Learning and Principles and Practice of Knowledge Dicovery in Databases (ECML-PKDD 2017), Skopje, Macedonia, Sep. 18, 2017, CEUR Workshop Proceedings Vol-1960, http://ceur-ws.org/Vol-1960/
- [68] E. Horvitz, B. Grosz, A. Greenwald, D. Parkes, C. Gomes, S. Smith, et al. (Eds.), Artificial Intelligence for the Social Good, Papers from the 2017 AAAI Spring Symposium, Technical Report SS-17-01, AAAI, Palo Alto, CA, 2017, https://www.aaai.org/Library/Symposia/Spring/ss17-01.php
- [69] J. Garriga, J. Piera, F. Bartumeus, A Bayesian framework for reputation in citizen science, In: Proceedings of the Second Workshop on Data Science for Social Good, CEUR Workshop Proceedings, 2017, 1960, 1-18, http://ceur-ws.org/Vol-1960/paper6.pdf
- [71] J. Auerbach, H. Barton, T. Blunt, V. Chaganti, B. Ghai, A. Meng, et al., Using data science as a community advocacy tool to promote equity in urban renewal programs: An analysis of Atlanta’s Anti-Displacement Tax Fund, In: Data Science for Social Good Conference 2017, Sep. 28-29, 2017, Chicago, IL, http://dssg.uchicago.edu/wp-content/uploads/2017/09/auerbach.pdf
- [73] M. Prasad, Back to the future: A framework for modelling altruistic intelligence explosions, In: AAAI 2017 Spring Symposium on Artificial Intelligence for the Social Good, https://aaai.org/ocs/index.php/SSS/SSS17/paper/view/15326
- [74] Y.-H. Wang, Y.-Y. Chen, S.-C. Chen, C.-K. Liu, T. C. Hsieh, Data for Social Good: A case study of building an effective public-private partnership on domestic violence prevention, In: Data Science for Social Good Conference 2017, Sep. 28- 29, 2017, Chicago, IL, http://dssg.uchicago.edu/wp-content/uploads/2017/09/hsieh.pdf
- [75] A. Delaunay, J. Guérin, Wandering detection within an embedded system for Alzheimer suffering patients, In: AAAI 2017 Spring Symposium on Artificial Intelligence for the Social Good, https://aaai.org/ocs/index.php/SSS/SSS17/paper/view/15317
- [76] O. Bendel, LADYBIRD: The animal-friendly robot vacuum cleaner, In: AAAI 2017 Spring Symposium on Artificial Intelligence for the Social Good, https://aaai.org/ocs/index.php/SSS/SSS17/paper/view/15277
- [77] P. Baumgartner, Challenges in assessing predictive bias, In: Data Science for Social Good Conference 2017, Sep. 28- 29, 2017, Chicago, IL, http://dssg.uchicago.edu/wp-content/uploads/2017/09/baumgartner.pdf
- [82] C.-K. Liu, T. C. Hsieh, Lessons learned from using data science to empower change agents across data silos, In: Data Science for Social Good Conference 2017, Sep. 28-29, 2017, Chicago, IL, https://dssg.uchicago.edu/wp-content/uploads/2017/09/liu.pdf
- [84] H. Holmestad, Predicting risk of long-term unemployment, In: Data Science for Social Good Conference 2017, Sep. 28-29, 2017, Chicago, IL, https://dssg.uchicago.edu/wp-content/uploads/2017/09/holmestad.pdf
Cited by
Del Río Castro, G., González Fernández, M. C., & Uruburu Colsa, Á. (2021) · Journal of Cleaner Production · doi:10.1016/j.jclepro.2020.122204 · Cited by 569
The future of AI and education: Some cautionary notes
Selwyn, N. (2022) · European Journal of Education · doi:10.1111/ejed.12532 · Cited by 401
Tackling climate change with machine learning
Rolnick, D., Donti, P. L., Kaack, L. H., Kochanski, K., Lacoste, A., Sankaran, K., Ross, A. S., Milojevic-Dupont, N., Jaques, N., Waldman-Brown, A., et al. (2023) · ACM Computing Surveys · doi:10.1145/3485128 · Cited by 336
“[5, 6]), ML for the developing world [7], and the computational sustainability”
BackgroundStahl, B. C. (2021) · SpringerBriefs in Research and Innovation Governance · doi:10.1007/978-3-030-69978-9 · Cited by 317
Ryan, M., & Stahl, B. C. (2021) · Journal of Information, Communication and Ethics in Society · doi:10.1108/jices-12-2019-0138 · Cited by 310
“As a starting point we used the ethical principles that are used by the EU ’ s HLEG (2019) as high-level coding points.”
BackgroundMethodInfluential
Metadata
- Authors
- Berendt, B.
- Affiliations
- KU Leuven (BE)
- Corresponding author
- Berendt, Bettina
- ORCID
- Berendt, Bettina
- Venue
- Paladyn, Journal of Behavioral Robotics
- Publisher
- Walter de Gruyter GmbH
- Published
- 2019-01-01
- Volume
- 10
- Issue
- 1
- Pages
- 44-65
- ISSN
- 2080-9778 (ISSN-L)
- 2081-4836
- OpenAlex
- W2898730196
- Cited by
- 8
- References
- 52
- Citations by year
- 2026: 1 · 2025: 1 · 2023: 3 · 2021: 1 · 2020: 2
- FWCI
- 1.31
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Locations
- Paladyn Journal of Behavioral Robotics (publishedVersion, cc-by)
- Lirias (acceptedVersion, other-oa)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
- Social Sciences › Sociology and Political Science › Innovation, Technology, and Society
- Physical Sciences › Computer Science › Artificial Intelligence › Artificial Intelligence Applications
- Keywords
- AI ethics, AI for social good, problem formulation, role of knowledge, AI design, moral norms
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.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
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
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: AI for the Common Good?! Pitfalls, challenges, and ethics pen-testing. https://app.cochat.ai/papers/10.1515/pjbr-2019-0004
- The paper
- Berendt, B. (2019). AI for the Common Good?! Pitfalls, challenges, and ethics pen-testing. Paladyn, Journal of Behavioral Robotics. https://doi.org/10.1515/pjbr-2019-0004
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.