Defining AI in Policy versus Practice
Open accessThe DOI matches its Crossref record; 22 of the 22 references it cites were found in a registry or an index.
Abstract
Recent concern about harms of information technologies motivate consideration of regulatory action to forestall or constrain certain developments in the field of artificial intelligence (AI). However, definitional ambiguity hampers the possibility of conversation about this urgent topic of public concern. Legal and regulatory interventions require agreed-upon definitions, but consensus around a definition of AI has been elusive, especially in policy conversations. With an eye towards practical working definitions and a broader understanding of positions on these issues, we survey experts and review published policy documents to examine researcher and policy-maker conceptions of AI. We find that while AI researchers favor definitions of AI that emphasize technical functionality, policy-makers instead use definitions that compare systems to human thinking and behavior. We point out that definitions adhering closely to the functionality of AI systems are more inclusive of technologies in use today, whereas definitions that emphasize human-like capabilities are most applicable to hypothetical future technologies. As a result of this gap, ethical and regulatory efforts may overemphasize concern about future technologies at the expense of pressing issues with existing deployed technologies.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
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?
In the citing paper
“In a recent study, researchers centre system and technical elements in their definitions, while policymakers focus on how systems compare to human thought or behaviour (Krafft et al., 2020).”
From Representation of Libraries in Artificial Intelligence Regulations and Implications for Ethics and Practice (2022) · cites this paper as Krafft et al., 2020
In the citing paper
“Krafft suggests that the OECD's policy-focused definition effectively reaches across the domains of research and policy and this definition also frames this paper”
From Representation of Libraries in Artificial Intelligence Regulations and Implications for Ethics and Practice (2022) · cites this paper as Krafft et al., 2020
In the citing paper
“This finding reflects a larger definitional disconnect affecting AI policy oversight efforts between researchers and policymakers in what technologies are considered to be AI, or are potential sources of algorithmic bias [41].”
From Toward situated interventions for algorithmic equity: lessons from the field (2020) · cites this paper as [41]
Citations
Cites
Floridi, L., Cowls, J., Beltrametti, M., Chatila, R., Chazerand, P., Dignum, V., Luetge, C., Madelin, R., Pagallo, U., Rossi, F., et al. (2018) · Minds and Machines · doi:10.1007/s11023-018-9482-5 · Cited by 4,774
Amazon Scraps Secret AI Recruiting Tool that Showed Bias against Women *
Dastin, J. (2022) · Ethics of Data and Analytics · doi:10.1201/9781003278290-44 · Cited by 1,165
How to Conduct a Mixed Methods Study: Recent Trends in a Rapidly Growing Literature
Small, M. L. (2011) · Annual Review of Sociology · doi:10.1146/annurev.soc.012809.102657 · Cited by 829 · cited 1×
“Our methodological motivation was complementary in combining different types of data [29].”
BackgroundFuture Progress in Artificial Intelligence: A Survey of Expert Opinion
Müller, V. C., & Bostrom, N. (2016) · Synthese Library · doi:10.1007/978-3-319-26485-1_33 · Cited by 655 · cited 1×
“In the long term, some believe AI technologies pose an existential risk to humanity by altering the scope of human agency and selfdetermination [24], or by the creation of autonomous weapons [3].”
BackgroundWhere fairness fails: data, algorithms, and the limits of antidiscrimination discourse
Hoffmann, A. L. (2019) · Information, Communication & Society · doi:10.1080/1369118x.2019.1573912 · Cited by 535 · cited 1×
“In the short term, evidence indicates that machine learning (ML) algorithms could contribute to oppression and discrimination due to historical legacies of injustice reflected in training data [2, 7, 13], directly to economic inequality through job displacement [14], or through a failure to reflectively account for wh…”
BackgroundSituating methods in the magic of Big Data and AI
Elish, M. C., & boyd, D. (2018) · Communication Monographs · doi:10.1080/03637751.2017.1375130 · Cited by 409 · cited 1×
“To address these concerns, the authors advocate for re-framing conversations around machine autonomy to foreground human actors and the broader sociotechnical context in which such systems are embedded, an injunction forcefully argued by other critical scholars as well [15].”
BackgroundAsaro, P. (2012) · International Review of the Red Cross · doi:10.1017/s1816383112000768 · Cited by 385 · cited 1×
“em [18]. In the long term, some believe AI technologies pose an existential risk to humanity by altering the scope of human agency and selfdetermination [24], or by the creation of autonomous weapons [3]. For many researchers in machine learning, the answer to these challenges has been technical, and there has been a…”
BackgroundThe IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems
Chatila, R., & Havens, J. C. (2019) · Intelligent Systems, Control and Automation: Science and Engineering · doi:10.1007/978-3-030-12524-0_2 · Cited by 307
Johnson, D. G., & Verdicchio, M. (2017) · Minds and Machines · doi:10.1007/s11023-017-9417-6 · Cited by 241 · cited 1×
“Johnson and Verdicchio [20] identify two primary concerns with the public reception of expert discourse on AI.”
BackgroundHow to do better with gender on surveys: a guide for HCI researchers
Spiel, K., Haimson, O. L., & Lottridge, D. (2019) · Interactions · doi:10.1145/3338283 · Cited by 207 · cited 1×
“ian or Pacific Islander, and 60% White. 4% entered their own race or ethnicity and 5% indicated they preferred not to say. In our second survey, which used a different phrasing of our gender question [31], 21% reported Woman as their gender, 77% Man, and 0% Non-binary. 0% entered their own gender identity and 2% indic…”
Fundamental Issues of Artificial Intelligence
(2016) · Synthese Library · doi:10.1007/978-3-319-26485-1 · Cited by 193
"Scary Robots": Examining Public Responses to AI
Cave, S., Coughlan, K., & Dihal, K. (2019) · Proceedings of the 2019 AAAI/ACM Conference on AI, Ethics, and Society · doi:10.1145/3306618.3314232 · Cited by 186 · cited 3×
“favor human thinking/behavior in their definitions of AI. One possible consequence of this is that policy documents may overly focus on the future social consequences of AI. As noted by Cave et al., [8, 10], policymaker conceptions of humanoid AI may focus on the retreating horizon of systems still-to-be-created at th…”
BackgroundInfluentialAn AI Race for Strategic Advantage: Rhetoric and Risks
Cave, S., & ÓhÉigeartaigh, S. S. (2018) · Proceedings of the 2018 AAAI/ACM Conference on AI, Ethics, and Society · doi:10.1145/3278721.3278780 · Cited by 124 · cited 1×
“, [8, 10], policymaker conceptions of humanoid AI may focus on the retreating horizon of systems still-to-be-created at the risk of passing over autonomous systems already in place.”
BackgroundAlgorithms for decision making
Riseth, A. N. (2018) · Oxford University Research Archive (ORA) (University of Oxford) · doi:10.5287/ora-rjb50ebd8 · Cited by 124
Fairness Through Computationally-Bounded Awareness
Kim, M. P., Reingold, O., & Rothblum, G. N. (2018) · arXiv · doi:10.48550/arxiv.1803.03239 · Cited by 97 · cited 1×
“ng, the answer to these challenges has been technical, and there has been a growth of work to promote computational approaches to fairness, accountability, and transparency in machine learning (e.g., [5, 21]). For others, these challenges may require regulatory intervention. In the regulatory line, widespread public c…”
BackgroundArtificial intelligence: past and future
Vardi, M. Y. (2011) · Communications of the ACM · doi:10.1145/2063176.2063177 · Cited by 61 · cited 1×
“Similarly, John McCarthy states, “As soon as it works, no one calls it AI anymore” [33].”
BackgroundMunicipal surveillance regulation and algorithmic accountability
Young, M., Katell, M., & Krafft, P. M. (2019) · Big Data & Society · doi:10.1177/2053951719868492 · Cited by 51 · cited 2×
“In choosing the six particular technologies we presented, we drew on prior work of ours that identified different conceptions of certain surveillance technologies among city stakeholders in Seattle [36].”
MethodSelf-citationOn preserving non-discrimination when combining expert advicePreprint
Blum, A. L., Gunasekar, S., Lykouris, T., & Srebro, N. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1810.11829 · Cited by 31 · cited 1×
“, [5, 21]).”
Inferring Work Task Automatability from AI Expert Evidence
Duckworth, P., Graham, L., & Osborne, M. A. (2019) · Proceedings of the AAAI/ACM Conference on AI Ethics and Society · doi:10.1145/3306618.3314247 · Cited by 17 · cited 1×
“In the short term, evidence indicates that machine learning (ML) algorithms could contribute to oppression and discrimination due to historical legacies of injustice reflected in training data [2, 7, 13], directly to economic inequality through job displacement [14], or through a failure to reflectively account for wh…”
BackgroundThat's AI? : a history and critique of the field
Sweeney, L. (1989) · doi:10.1184/r1/6610337 · Cited by 8
Municipal Surveillance Regulation and Algorithmic AccountabilityPreprint
Young, M., Katell, M. A., & Krafft, P. M. (2019) · SocArXiv (OSF Preprints) · doi:10.31235/osf.io/zx2sw · Cited by 7
Self-citationLiu, Y., & Shen, V. (2012) · Communications of the ACM · doi:10.1145/2133806.2133807 · Cited by 2
Research works printed but not linked: 16
- Philip E Agre. 1997. Lessons learned in trying to reform AI . Social Science, Technical Systems, and Cooperative Work: Beyond the Great Divide (1997), 131.
- Julia Angwin, Jeff Larson, Surya Mattu, and Lauren Kirchner. 2016. Machine bias. ProPublica, May , Vol. 23 (2016).
- Joanna J Bryson. 2019. The Past Decade and Future of AI's Impact on Society. Towards a New Enlightenment: A Transcendent Decade. Madrid: Turner-BVVA. (2019).
- Joy Buolamwini and Timnit Gebru. 2018. Gender shades: Intersectional accuracy disparities in commercial gender classification. In Conference on Fairness, Accountability and Transparency. 77--91.
- Stephen Cave, Claire Craig, Kanta Sarasvati Dihal, Sarah Dillon, Jessica Montgomery, Beth Singler, and Lindsay Taylor. 2018. Portrayals and perceptions of AI and why they matter . Technical Report. The Royal Society.
- Ariel Conn. 2016. The problem of defining autonomous weapons. The Future of Life Institute (2016).
- 10.5555/3298239.3298381 · doi:10.5555/3298239.3298381
- Bill Howe. 2019. Protect the public from bias in automated decision systems. Seattle Times (2019).
- Donald Michie. 1971. Formation and execution of plans by machine .University of Edinburgh Department of Machine Intelligence & Perception.
- Brent Mittelstadt. 2019. AI ethics--Too principled to fail? arXiv preprint arXiv:1906.06668 (2019). · arXiv
- Group of Governmental Experts. 2019. Group of Governmental Experts on Lethal Autonomous Weapons. Fifth Review Conference of the High Contracting Parties to the Convention on Certain Conventional Weapons (2019).
- Stuart J Russell and Peter Norvig. 2016. Artificial Intelligence: A Modern Approach .Malaysia; Pearson Education Limited,.
- Science and Technology Committee. 2018. Algorithms in decision-making. UK Parliament House of Commons , Vol. 36 (2018).
- Brad Smith and Harry Shum. 2018. The future computed: Artificial intelligence and its role in society. Microsoft Corporation (2018).
- Sarah Myers West, Meredith Whittaker, and Kate Crawford. 2019. Discriminating Systems: Gender, Race, and Power in AI. Technical Report. AI Now Institute.
- Meredith Whittaker, Kate Crawford, Roel Dobbe, Genevieve Fried, Elizabeth Kaziunas, Mathur Varoon, Sarah Myers West, Rashida Richardson, Jason Schultz, and Oscar Schwartz. 2018. AI Now 2018 Report . Technical Report. AI Now Institute, New York, NY .
Cited by
Trust in AI and Its Role in the Acceptance of AI Technologies
Choung, H., David, P., & Ross, A. (2023) · International Journal of Human–Computer Interaction · doi:10.1080/10447318.2022.2050543 · Cited by 963
“The uniqueness of AI is characterized by the integration of its functionality with humanlike capabilities (Krafft et al., 2020).”
BackgroundDiscourses of artificial intelligence in higher education: a critical literature review
Bearman, M., Ryan, J., & Ajjawi, R. (2023) · Higher Education · doi:10.1007/s10734-022-00937-2 · Cited by 545
“AI may, simply, mean different things to different people (Bear-man & Luckin, 2020; Krafft et al., 2020).”
Backgroundvan Noordt, C., & Misuraca, G. (2022) · Government Information Quarterly · doi:10.1016/j.giq.2022.101714 · Cited by 356
Out of the laboratory and into the classroom: the future of artificial intelligence in education
Schiff, D. (2021) · AI & SOCIETY · doi:10.1007/s00146-020-01033-8 · Cited by 334
Rakova, B., Yang, J., Cramer, H., & Chowdhury, R. (2021) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3449081 · Cited by 329
“[18] point out that even between experts, there is a disconnect between policymakers and researchers’ definitions of such foundational terms as ‘AI’.”
Background
Metadata
- Authors
- Krafft, P. M., Young, M., Katell, M., Huang, K., & Bugingo, G.
- Affiliations
- University of Oxford (GB)
- University of Washington (US)
- Harvard University (US)
- Venue
- Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society
- Publisher
- ACM
- Published
- 2020-02-07
- Pages
- 72-78
- ISSN
- 3065-8365 (ISSN-L)
- OpenAlex
- W3004528080
- Cited by
- 129
- References
- 20
- Citations by year
- 2026: 15 · 2025: 22 · 2024: 24 · 2023: 21 · 2022: 19 · 2021: 18 · 2020: 10
- FWCI
- 17.64
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- Proceedings of the AAAI/ACM Conference on AI Ethics and Society (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- Gordon and Betty Moore Foundation (2013-10-29)
- Washington Research Foundation
- Alfred P. Sloan Foundation (3835)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
- Social Sciences › Law › Digitalization, Law, and Regulation
- Physical Sciences › Computer Science › Artificial Intelligence › Artificial Intelligence Applications
- Keywords
- AI policy, AI regulation, expert survey, AI ethics, regulatory overreach
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
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.