Pith. sign in

REVIEW 3 major objections 6 minor 50 references

When AI Becomes Routine: A Decade of Public AI Mediation in Korean Go Commentary

T0 review · 3 major / 6 minor · reviewed 2026-07-31 · grok-4.5

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Once AI is routine in Korean Go commentary, the word “AI” recedes while winrate talk stays—and that recession is the signature of domestication.

desk verdict Careful decade-scale measurement of how AI talk recedes into interface metrics in Korean Go commentary; the descriptive shift holds, the domestication reading is scoped but underdetermined. read the letter →

arxiv 2607.28332 v1 pith:VHSL4ERL submitted 2026-07-30 cs.CY

Haewoon Kwak This is my paper · ORCID
classification cs.CY
keywords AImediationdomesticationcontestabilityGocommentarysourceattributioninterfacerenderingexpertintermediariesethics
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

After superhuman Go engines became everyday tools, this paper asks how experts make machine judgment publicly intelligible and attributable. Across roughly 1,900 hours of Korean YouTube commentary from 2016–2025, winrate graphs sit on screen nearly all the time in late institutional broadcasts, yet only a few percent of sentences overtly activate AI. What fades is the source label, not the metric: commentators still read winrates and point gaps, but stop saying “AI.” The author reads that pattern as domestication’s communicative signature, shows creator channels lean further into interface-only talk than institutional ones, and offers a typology of source-foregrounding versus source-receding mediation. Those forms leave audiences different hooks for recognizing and questioning the machine—stakes that rise wherever AI is less trustworthy than in Go.

What carries the argument

A precision-first keyword-anchored AI-salient subset (explicit AI naming plus interface/metric language), used to track composition over four phases and across institutional versus creator channels, plus a mediation-form typology that splits source-foregrounding practices (naming, human–machine contrast, metric translation) from source-receding ones (interface rendering, reportive relay).

What would settle it

Compare matched late-routine segments where the winrate graph is deliberately hidden or removed: if explicit “AI” naming does not rebound when the graphic is gone, the domestication reading is supported; if naming surges only to replace the missing graphic, ordinary broadcast redundancy-avoidance is the better account.

Watch

Extended reading notes

Core claim

In late-routine Korean Go commentary, visual AI is near-saturated while verbal AI-salient talk stays small; inside that talk, explicit naming yields to interface and metric rendering (more so on creator channels than institutional ones). The recession of the source label “AI,” while metric talk persists, is the communicative signature of domestication, and source-foregrounding versus source-receding forms preserve different audience hooks of contestability.

Load-bearing premise

That commentators stop saying “AI” mainly because machine judgment has been socially domesticated, not simply because they avoid verbally narrating a graphic that is already always on screen.

Editorial extensions

If this is right

  • Public contestability of AI judgment depends on mediation form, not only on model outputs or internal explanations.
  • Creator-style solo narration will tend toward source-receding interface talk more than multi-commentator institutional formats.
  • In low-stakes, near-oracle domains, naturalizing the source can be functional fluency rather than a transparency failure.
  • Where AI is less reliable or stakes are higher, the same compositional shift toward interface-only talk raises governance concern about eroded discursive anchors for challenge.
  • Auditing deployed AI should include how intermediaries name, translate, or fold away the machine source in public speech.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Platforms and vendors optimizing for fluent uptake have a structural incentive to push interfaces and scripts toward source-receding forms—the paper’s adverse-impact note implies an audit-versus-optimization fork.
  • A controlled A/B on-air test (graph on vs graph off, same commentators) would cleanly separate domestication from UI redundancy and travel beyond Go.
  • Similar source-label recession may already be measurable in sports betting overlays, medical dashboard readouts, or credit-score call centers once the metric is permanently visible.
  • Preserving occasional marked naming—even when the interface is saturated—may be a cheap contestability intervention without blocking routine use.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. The paper studies how superhuman Go AI is publicly mediated in Korean YouTube commentary over 2016–2025 (~1,900 hours; institutional and creator channels). Using a precision-first keyword-anchored AI-salient subset, OCR winrate-visibility checks, and phase/denominator robustness, it documents a late-period visual–verbal asymmetry (winrate graphs ~98% of institutional broadcast time vs. 2.63% AI-salient sentences) and a compositional shift inside that subset from explicit source naming toward interface/metric rendering, stronger in creator than institutional talk. It interprets the recession of the source label (while metric talk persists) as the communicative signature of domestication, offers a source-foregrounding vs. source-receding mediation typology tied to hooks of contestability, and scopes cross-domain governance implications as untested transfer predictions rather than demonstrated outcomes.

Significance. If the descriptive results hold—as the audits, OCR calibration (including a Phase 3-specific detector), boundary/denominator checks, and video-clustered GEE largely support—this is a rare decade-scale naturalistic account of public AI mediation after routine superhuman adoption, not merely anticipation or short-run reaction. Strengths include a conservative, audited measurement design; released code and keyword rules; explicit separation of volume vs. composition; and a usable typology that connects mediation form to contestability without overclaiming legitimacy outcomes. For AIES/cs.CY, the contribution is a concrete post-deployment vocabulary for how intermediaries make machine judgment attributable (or unmarked) under different media logics.

major comments (3)
  1. [Abstract; §6.1; §6.6] §6.1 correctly states that verbal-only data cannot discriminate socio-cognitive domestication from ordinary broadcast redundancy-avoidance against a near-saturated graphic. The abstract and §6.6 still lead with “we read this recession as the communicative signature of domestication” without that caveat. Because the governance transfer (§6.3) depends on the stronger reading, the abstract/conclusion should foreground the underdetermination and state the primary result as the measured visual–verbal asymmetry plus the naming→interface compositional shift, with domestication as one interpretive frame among those the data cannot fully separate.
  2. [§5.2; §6.4] §5.2 reports creator vs. institutional composition with GEE cluster=video (OR=0.66, p=0.010) but notes cluster=channel (K=7) is non-significant (p=0.37), and explicitly says no field-level claim about mediation norms is supported. Later prose (§6.4, conclusion) still generalizes “institutional formats… retain more marked moments than creator formats.” Keep every field-level claim strictly at the video-within-corpus level already justified, or add channel-robust evidence; do not let the media-logic narrative outrun the clustering caveat.
  3. [§4.1; §5.1; §6.1] The anti-suppression argument (§4.1; §6.1)—that winrate/point-gap talk persists while only the label “AI” recedes—is load-bearing for preferring naturalization over simple on-screen-scoreboard silence. Persistence is shown qualitatively and via interface-marker shares inside the AI-salient subset, but not as a longitudinal rate of metric talk conditional on graph visibility outside that subset. A small supplementary contrast (e.g., rate of bare winrate/point-gap sentences that fail the strict AI-salient rule, or timed graph-movement narrations with vs. without source labels in Phase 3 vs. 4) would make this hinge claim less interpretive.
minor comments (6)
  1. [Table 3; §5.1] Table 3 Phase 4 BadukTV lists Explicit 53.67 and Interface 49.86 (can exceed 100% with overlap), while the body gives exclusive partitions (50.1/46.3/3.5). Add a note that marker columns are non-exclusive and point readers to the exclusive breakdown.
  2. [Figure 2] Figure 2 scales per-minute ×10; state the scaling explicitly in the caption (it is only in the body) to avoid misreading absolute levels.
  3. [Table 3; Methods] Frictional uptake is appropriately demoted after low marker-level κ (§Methods; Appendix F). Consider moving it out of the main summary table or marking it “descriptive only” in the table header so readers do not treat it as parallel to naming/interface.
  4. [§3; §5.1] D_Long selection on high-visibility tournament videos is well motivated but should be restated once when interpreting absolute AI-salient shares (2.63%) so readers do not treat them as population rates for all Korean Go video.
  5. [§3; Table 1] Minor consistency: corpus “approximately 1,900 hours” vs. D_Long “~1,394 hours” plus other sets—give a one-line hour breakdown by dataset in §3 or Table 1.
  6. [Table 1; Appendix A] Appendix claim-to-dataset map (Table S1) is helpful; a shortened version in the main text near Table 1 would aid readers who skip the supplement.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: observational keyword measures and an interpretive domestication frame, not a closed derivation.

full rationale

This is a longitudinal observational study of Korean Go commentary. Its load-bearing empirical claims (AI-salient sentence shares rising across phases; Phase-4 composition shifting toward interface-only mediation, more so on creator channels; ~98% winrate-graph visibility vs 2.63% verbal AI-salient talk) are computed from audited keyword rules, OCR visibility checks, and denominator/boundary robustness tests. Those quantities are not fitted to force the domestication conclusion, nor defined in terms of it. Domestication and the source-foregrounding/source-receding typology are interpretive frames applied after the counts; the paper explicitly treats the socio-cognitive vs broadcast-UI-redundancy reading as underdetermined by verbal data alone (§6.1). External citations (Silverstone/Haddon domestication, contestability/XAI literature) supply vocabulary, not a self-citation uniqueness chain. No fitted-input-as-prediction, self-definitional loop, or renaming of a known result as a first-principles derivation appears. Score 0 is the honest finding.

Assumptions & free parameters 3 free parameters · 6 assumptions · 4 invented entities

Load-bearing structure is empirical and interpretive, not axiomatic-mathematical. The central claim rests on domain assumptions about what keyword hits mean, on phase periodization choices, on treating high-visibility tournament commentary as the estimand, and on reading source-label recession as domestication. No fitted physical constants; free parameters are analytic design choices (phase cut, keyword tiers). Invented entities are conceptual constructs (typology cells, hooks of contestability), not ontological posits.

free parameters (3)
  • Phase 3/4 boundary (Jan 2021) = 2021-01-01
    Analytic cut defining “routine integration”; paper shows robustness across Jan 2020–Jan 2024 but the main reported Phase 4 figures use this hand-chosen boundary.
  • Strict AI-salient keyword inventory and exclusion rules = precision-first lexicon (Table 2); 96.7% audited precision
    Which tokens count as explicit AI vs interface/metric, and which bare winrate/percentage contexts are excluded, directly determine the 2.63% rate and composition shares.
  • D_Long stratified sample (400 high-visibility videos) = 400 videos / ~1394 hours
    Selection on viewership and major tournaments defines the estimand; average views 80k vs 11k unselected. Composition claims are conditional on this sample design.
assumptions (6)
  • ad hoc to paper Explicit AI naming is a marked surface form whose declining share relative to interface-only mediation indexes domestication of machine judgment in public talk.
    Core interpretive bridge from counts to “communicative signature of domestication” (§4.1, §6.1); motivated by Silverstone/Haddon domestication theory but not entailed by it.
  • domain assumption In Phases 3–4, on-screen winrate/recommendation layers already shape commentator reasoning even when AI is not named, so verbal AI-salient share is not a measure of total AI dependence.
    Stated in §4.1; supported by OCR visibility checks but still an assumption about silent uptake.
  • ad hoc to paper Persistence of verbal winrate/point-gap talk while the label “AI” recedes rules out simple on-screen-scoreboard suppression as the full explanation.
    Key anti-alternative argument in §4.1; necessary for the naturalization reading over redundancy-avoidance.
  • domain assumption Source-foregrounding forms preserve discursive “hooks of contestability”; source-receding forms erode them.
    Normative-analytic link used for AIES transfer (§6.2–6.3); grounded in contestable-AI literature but not directly measured as audience challenge rates (chat evidence is sparse/exploratory).
  • domain assumption Whisper Base keyword capture is adequate for the AI-anchor terms that enter the backbone.
    Methods: 120-row STT audit with 100% AI-anchor preservation; still assumes accents/fast creator speech do not systematically hide naming.
  • domain assumption High-visibility live tournament commentary is the right estimand for public mediation of AI judgment in Korean Go.
    §3 video selection; excludes average Go video and private review rooms.
invented entities (4)
  • AI-salient subset (strict keyword-anchored) independent evidence
    purpose: Operational lower bound on overt verbal AI activation; backbone for all volume and composition claims.
    Constructed for this paper; validated by precision/FN audits but definitionally tied to the authors’ lexicon.
  • Source-foregrounding vs source-receding mediation typology (five forms)
    purpose: Organize micro-patterns and link mediation format to contestability hooks for cross-domain transfer.
    Derived from qualitative inspection of the AI-salient subset (§5.5, Table 4); not independently measured outside this corpus.
  • Hooks of contestability
    purpose: Name the discursive anchors (e.g., the word “AI”) through which audiences can recognize and challenge machine source.
    Conceptual coinage bridging mediation form to AIES contestability literature; audience chat only illustratively supports it.
  • Routine superhuman AI (social condition)
    purpose: Frame the post-adoption setting the paper studies.
    Definitional framing in §1; useful but not an empirical discovery.

how reviews work

0 comments
Cite this review

Pith. "Pith review of When AI Becomes Routine: A Decade of Public AI Mediation in Korean Go Commentary." pith.science (2026). https://pith.science/paper/VHSL4ERL

@misc{pith2026260728332,
  author       = {Pith},
  title        = {Pith review of: When AI Becomes Routine: A Decade of Public AI Mediation in Korean Go Commentary},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VHSL4ERL}},
  note         = {Machine review of arXiv:2607.28332}
}
abstract

When AI systems surpass elite human performance and settle into everyday expert practice, the question that follows is how machine judgment is made publicly intelligible and attributable. We study Korean Go commentary on YouTube, where AI systems such as KataGo became standard analytic tools after AlphaGo. Our corpus spans a decade (2016--2025) and approximately $1{,}900$ hours of footage across institutional broadcasters and creator-led channels, in four phases of AI availability. We document a widening asymmetry between visual and verbal AI presence: AI winrate graphs are visible for about $98\%$ of late-period institutional broadcast time, yet AI-salient talk accounts for only $2.63\%$ of sentences. What recedes is the source label, not the metric: winrate and point-gap talk persists while ``AI'' itself goes unsaid. We read this recession as the communicative signature of domestication. Our strongest evidence is a compositional shift in verbal mediation: explicit naming gives way to interface rendering, and creator-led commentary leans further toward it than institutional commentary. We develop a typology distinguishing source-foregrounding from source-receding mediation, and argue that the two preserve different hooks of contestability: discursive anchors through which audiences can recognize and question the machine source. The stakes of that difference rise in domains where AI is less reliable than in Go.

Figures

Figures reproduced from arXiv: 2607.28332 by the authors.

Figure 1
Figure 1. Typical broadcast layouts. (a) Institutional simul [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. BadukTV longitudinal AI-salient share across [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Phase 4 channel-level composition of the AI [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

50 extracted references · 2 linked inside Pith

  1. [1]

    Alfrink, K.; Keller, I.; Kortuem, G.; and Doorn, N. 2023. Contestable AI by Design: Towards a Framework. Minds and Machines, 33: 613--639

    2023

  2. [2]

    Almada, M. 2019. Human Intervention in Automated Decision-Making: Toward the Construction of Contestable Systems. In Proceedings of the Seventeenth International Conference on Artificial Intelligence and Law, 2--11. ACM

    2019

  3. [3]

    Angwin, J.; Larson, J.; Mattu, S.; and Kirchner, L. 2022. Machine bias. In Ethics of data and analytics, 254--264. Auerbach Publications

    2022

  4. [4]

    BadukTV . 2026. BadukTV YouTube Channel. https://www.youtube.com/@baduk_tv. Accessed March 29, 2026

    2026

  5. [5]

    Berker, T.; Hartmann, M.; Punie, Y.; and Ward, K. J., eds. 2006. Domestication of Media and Technology. Maidenhead: Open University Press

    2006

  6. [6]

    L.; Vervier, L.; and Ziefle, M

    Brauner, P.; Glawe, F.; Liehner, G. L.; Vervier, L.; and Ziefle, M. 2025. Mapping public perception of artificial intelligence: Expectations, risk--benefit tradeoffs, and value as determinants for societal acceptance. Technological Forecasting and Social Change, 220: 124304. Doi:10.1016/j.techfore.2025.124304

    arXiv 2025

  7. [7]

    J.; Winter, S.; Steiner, D.; Wilcox, L.; and Terry, M

    Cai, C. J.; Winter, S.; Steiner, D.; Wilcox, L.; and Terry, M. 2019. ``Hello AI'': Uncovering the Onboarding Needs of Medical Practitioners for Human-AI Collaborative Decision-Making. Proceedings of the ACM on Human-Computer Interaction, 3(CSCW): 1--24

    2019

  8. [8]

    Castleman, J.; and Korolova, A. 2024. Why Am I Still Seeing This: Measuring the Effectiveness of Ad Controls and Explanations in AI -Mediated Ad Targeting Systems. In Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society, volume 7, 255--266. AAAI Press

    2024

Show all 50 references
  1. [9]

    ChoHyeyeon . 2026. Cho Hyeyeon Pro 9-dan YouTube Channel. https://www.youtube.com/channel/UCaVMWUQcMRqNxNrsBhhS85A. Accessed July 30, 2026

    2026

  2. [10]

    J.; Simmons, J

    Dietvorst, B. J.; Simmons, J. P.; and Massey, C. 2015. Algorithm aversion: people erroneously avoid algorithms after seeing them err. Journal of experimental psychology: General, 144(1): 114

    2015

  3. [11]

    DongneBaduk . 2026. DongneBaduk YouTube Channel. https://www.youtube.com/@ Accessed July 30, 2026

    2026

  4. [12]

    Doshi-Velez, F.; and Kim, B. 2017. Towards a rigorous science of interpretable machine learning. arXiv preprint arXiv:1702.08608

    2017 arXiv

  5. [13]

    EBS Documentary . 2026. 바둑계는 끝났다, 신의 경지까지 올라간 인공지능에 절대 못 이긴다. 알파고 대국 10년 후 | 다큐프라임 | \#골라듄다큐. https://www.youtube.com/watch?v=gz2Ig7Jf88I. Uploaded March 11, 2026. Accessed April 18, 2026

    2026

  6. [14]

    V.; Muller, M.; Riedl, M

    Ehsan, U.; Liao, Q. V.; Muller, M.; Riedl, M. O.; and Weisz, J. 2021. Expanding Explainability: Towards Social Transparency in AI Systems. In Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems, 1--19. ACM

    2021

  7. [15]

    Ensmenger, N. 2012. Is chess the drosophila of artificial intelligence? A social history of an algorithm. Social Studies of Science, 42(1): 5--30

    2012

  8. [16]

    G.; and Rahwan, I

    Epstein, Z.; Levine, S.; Rand, D. G.; and Rahwan, I. 2020. Who gets credit for AI-generated art? iScience, 23(9)

    2020

  9. [17]

    Gelly, S.; Kocsis, L.; Schoenauer, M.; Sebag, M.; Silver, D.; Szepesv \'a ri, C.; and Teytaud, O. 2012. The grand challenge of computer Go: Monte Carlo tree search and extensions. Communications of the ACM, 55(3): 106--113

    2012

  10. [18]

    Gibson, J. J. 1979. The ecological approach to visual perception. Houghton Mifflin

    1979

  11. [19]

    Glikson, E.; and Woolley, A. W. 2020. Human trust in artificial intelligence: Review of empirical research. Academy of management annals, 14(2): 627--660

    2020

  12. [20]

    Green, B.; and Chen, Y. 2019. Disparate Interactions: An Algorithm-in-the-Loop Analysis of Fairness in Risk Assessments. In Proceedings of the Conference on Fairness, Accountability, and Transparency, 90--99

    2019

  13. [21]

    Hartmann, M. 2013. From domestication to mediated mobilism. Mobile Media & Communication, 1(1): 42--49

    2013

  14. [22]

    Hjarvard, S. 2008. The Mediatization of Society: A Theory of the Media as Agents of Social and Cultural Change. Nordicom Review, 29(2): 105--134

    2008

  15. [23]

    Hutchby, I. 2001. Technologies, texts and affordances. Sociology, 35(2): 441--456

    2001

  16. [24]

    K-Baduk . 2026. K-Baduk YouTube Channel. https://www.youtube.com/@kbaduktv. Accessed March 29, 2026

    2026

  17. [25]

    Kaur, H.; Nori, H.; Jenkins, S.; Caruana, R.; Wallach, H.; and Vaughan, J. W. 2020. Interpreting Interpretability: Understanding Data Scientists' Use of Interpretability Tools for Machine Learning. In Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems...

    2020

  18. [26]

    Korea Baduk Association . 2020. 10대 핫이슈로 돌아본 2020년 바둑TV! https://m.baduk.or.kr/news/B02_view.asp?news_sum_no=7538. BadukTV news feature on 2020 broadcast-screen changes, including AI position evaluation standardization around KataGo. Published October 27, 2020. Accessed April 18, 2026

    2020

  19. [27]

    Korea Baduk Association . 2021. 핫피플 | 한국바둑AI연구소 이현호 대표. https://m.baduk.or.kr/news/B03_view.asp?news_no=155. 월간바둑, published July 2, 2021. Accessed April 18, 2026

    2021

  20. [28]

    Korea Baduk Association . 2024. 김만수, 초반 공부 끝판왕 ‘초반 70수’ 출간. https://m.baduk.or.kr/news/B01_view.asp?news_no=4940. Published April 18, 2024. Accessed April 18, 2026

    2024

  21. [29]

    Lebovitz, S.; Lifshitz-Assaf, H.; and Levina, N. 2022. To engage or not to engage with AI for critical judgments: How professionals deal with opacity when using AI for medical diagnosis. Organization science, 33(1): 126--148

    2022

  22. [30]

    Lee, M.-c. 2024. Kiwi: Developing a Korean morphological analyzer based on statistical language models and skip-bigram. Korean Journal of Digital Humanities, 1(1): 109--136

    2024

  23. [31]

    LeeHyunWookTV . 2026. LeeHyunWookTV YouTube Channel. https://www.youtube.com/@leehyunwooktv. Accessed March 29, 2026

    2026

  24. [32]

    V.; Gruen, D.; and Miller, S

    Liao, Q. V.; Gruen, D.; and Miller, S. 2020. Questioning the AI: Informing Design Practices for Explainable AI User Experiences. In Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems, 1--15. ACM

    2020

  25. [33]

    M.; Minson, J

    Logg, J. M.; Minson, J. A.; and Moore, D. A. 2019. Algorithm appreciation: People prefer algorithmic to human judgment. Organizational Behavior and Human Decision Processes, 151: 90--103

    2019

  26. [34]

    Lyons, H.; Velloso, E.; and Miller, T. 2021. Conceptualising Contestability: Perspectives on Contesting Algorithmic Decisions. Proceedings of the ACM on Human-Computer Interaction, 5(CSCW1): 1--22

    2021

  27. [35]

    V.; Omar, Z

    Menon, A. V.; Omar, Z. A.; Nahar, N.; Papademetris, X.; Fiellin, L. E.; and K \"a stner, C. 2024. Lessons from Clinical Communications for Explainable AI . In Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society, volume 7, 958--970. AAAI Press

    2024

  28. [36]

    A.; Singh, R.; and Elish, M

    Metcalf, J.; Moss, E.; Watkins, E. A.; Singh, R.; and Elish, M. C. 2021. Algorithmic Impact Assessments and Accountability: The Co-construction of Impacts. In Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency, 735--746. ACM

    2021

  29. [37]

    Miller, T. 2019. Explanation in artificial intelligence: Insights from the social sciences. Artificial Intelligence, 267: 1--38

    2019

  30. [38]

    Miyazaki, K.; Murayama, T.; Uchiba, T.; An, J.; and Kwak, H. 2024. Public perception of generative AI on Twitter: an empirical study based on occupation and usage. EPJ Data Science, 13(1): 2

    2024

  31. [39]

    ProYeonwoo . 2026. ProYeonwoo YouTube Channel. https://www.youtube.com/@proyeonwoo. Accessed March 29, 2026

    2026

  32. [40]

    W.; Xu, T.; Brockman, G.; McLeavey, C.; and Sutskever, I

    Radford, A.; Kim, J. W.; Xu, T.; Brockman, G.; McLeavey, C.; and Sutskever, I. 2023. Robust speech recognition via large-scale weak supervision. In International conference on machine learning, 28492--28518. PMLR

    2023

  33. [41]

    Rahwan, I. 2018. Society-in-the-Loop: Programming the Algorithmic Social Contract. Ethics and Information Technology, 20(1): 5--14

    2018

  34. [42]

    D.; Smart, A.; White, R

    Raji, I. D.; Smart, A.; White, R. N.; Mitchell, M.; Gebru, T.; Hutchinson, B.; Smith-Loud, J.; Theron, D.; and Barnes, P. 2020. Closing the AI accountability gap: Defining an end-to-end framework for internal algorithmic auditing. In Proceedings of the 2020 conference on fairn...

    2020

  35. [43]

    RyuSihunWorld . 2026. Ryu Shikun's Go World YouTube Channel. https://www.youtube.com/channel/UCqcHLpDSKKSkQDg1xO0B4EA. Accessed July 30, 2026

    2026

  36. [44]

    Shneiderman, B. 2020. Human-Centered Artificial Intelligence: Reliable, Safe & Trustworthy. International Journal of Human--Computer Interaction, 36(6): 495--504

    2020

  37. [45]

    J.; and Vos, T

    Shoemaker, P. J.; and Vos, T. P. 2009. Gatekeeping Theory. New York: Routledge

    2009

  38. [46]

    J.; Guez, A.; Sifre, L.; Van Den Driessche, G.; Schrittwieser, J.; Antonoglou, I.; Panneershelvam, V.; Lanctot, M.; et al

    Silver, D.; Huang, A.; Maddison, C. J.; Guez, A.; Sifre, L.; Van Den Driessche, G.; Schrittwieser, J.; Antonoglou, I.; Panneershelvam, V.; Lanctot, M.; et al. 2016. Mastering the game of Go with deep neural networks and tree search. nature, 529(7587): 484--489

    2016

  39. [47]

    Silverstone, R.; and Haddon, L. 1996. Design and the domestication of information and communication technologies: Technical change and everyday life. In Communication by design, 44--74. Oxford University Press

    1996

  40. [48]

    J.; Mosier, K

    Skitka, L. J.; Mosier, K. L.; and Burdick, M. 1999. Does automation bias decision-making? International Journal of Human-Computer Studies, 51(5): 991--1006

    1999

  41. [49]

    Stark, L.; and Hoffmann, A. L. 2019. Data is the New What? Popular Metaphors and Professional Ethics in Emerging Data Culture. Journal of Cultural Analytics, 4(1)

    2019

  42. [50]

    Wu, D. J. 2019. Accelerating self-play learning in Go. arXiv preprint arXiv:1902.10565

    2019 arXiv

Pith tools

Reviewed July 31, 2026 · model on record in the stance chip above.