Pith. sign in

REVIEW 5 major objections 7 minor 154 references

Global Perspectives of AI Risks and Harms: Analyzing the Negative Impacts of AI Technologies as Prioritized by News Media

T0 review · 5 major / 7 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read Global news coverage of AI skews toward societal risks, study of 42,853 articles finds.

desk verdict Useful global map of AI risk coverage in English-language news, but the headline prevalence ranking rests on an unverified 26.2% scrape sample and a partial taxonomy. read the letter →

arxiv 2501.14040 v2 pith:N34OTL7P submitted 2025-01-23 cs.CY

Mowafak Allaham , Kimon Kieslich , Nicholas Diakopoulos This is my paper · ORCID
classification cs.CY
keywords AIrisksnewsmediacoverageriskprioritizationsocietalpoliticalbiasgovernanceGPT-4oannotationglobalanalysis
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

The paper tries to establish that, when global English-language news coverage of AI is read as a record of public concerns, one skew shows up consistently: societal risks dominate, then legal and rights-related risks, with cognitive, content-safety, existential, and environmental risks trailing far behind. The authors analyze 42,853 articles from 27 countries across six regions and argue that this ordering reflects how media set the agenda for what publics and policymakers treat as important. They extend an existing AI risk taxonomy with 16 newly observed categories and show that the political leaning of an outlet changes which risks get reported. A reader should care because this offers a bottom-up, geographically inclusive source of evidence for AI risk assessment that is not limited to expert or corporate perspective.

What carries the argument

The central object is the emerging six-category risk scheme and the annotation pipeline that produces it. An LLM summarizes negative impacts from each article, yielding 47,731 impact statements; two human annotators label 1,060 of them against the AIR-taxonomy's Level-2 categories plus new categories as they emerge, reaching an inter-coder reliability of 0.97; the same LLM then classifies the full set, and the six categories with recall of at least 0.70 are kept for prevalence reporting. The AIR-taxonomy is a risk categorization distilled from government regulations and company policies, and the paper uses it as the baseline it extends with 16 novel categories proposed as eight additions at Level-1.

What would settle it

If a complete corpus that includes paywalled and non-English articles from the same 27 countries placed a different category at the top or lifted environmental or existential coverage above the 1.9% and 7.2% shares reported here, the paper's claim of a global societal-risk skew would fall.

Watch

Extended reading notes

Core claim

The paper claims that global news coverage of AI, as measured in a sample of 42,853 English-language articles from 27 countries, prioritizes Societal Risks (50.6% of 16,312 articles with negative impacts), followed by Legal & Rights-related Risks (32.9%), Cognitive Risks (14.2%), Content Safety Risks (8.3%), Existential Risks (7.2%), and Environmental Risks (1.9%). It further claims that this skew holds across all six regions, with societal risks the most prevalent category everywhere, and that political bias shapes coverage: fringe media report more on risks than least-biased outlets, right-biased media lead on existential risks and underreport environmental ones, and left-biased media emphasize legal and rights-related issues. The paper also presents 16 risk categories absent from the AIR-taxonomy and proposes grouping them into eight additions at the top level, arguing that expert-driven taxonomies omit risks that are salient in journalistic coverage.

Load-bearing premise

The 42,853 scraped articles, which are 26.2% of the eligible links and are dominated by the United States, are assumed to represent the AI risk coverage of all English-language national news in the 27 countries.

Editorial extensions

If this is right

  • If the prevalence skew holds, stakeholders using media as a proxy for public concern will see societal and legal/rights harms as the issues publics most strongly associate with AI.
  • The very low share of environmental coverage (1.9%) suggests environmental costs are being underweighted in public discourse relative to their physical scale, a gap regulators and companies could address.
  • The finding that left- and right-fringe outlets report risks more than least-biased ones implies that AI risk debates are partly polarized, so governance efforts should anticipate political framing.
  • Regional differences in rank order, such as content-safety risks ranking third in the Middle East and legal/rights risks peaking in Oceania, imply that global risk assessments need regional calibration rather than a single priority list.

Reading between the lines

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

  • The published percentages describe only the 26.2% of URLs that could be scraped; a replication covering paywalled and non-English national outlets would test whether the global ranking survives.
  • The six chosen categories were selected partly on LLM recall, so categories excluded for annotation quality, such as governance and information risks, may still be central to public discourse and merit targeted study.
  • The word 'prioritization' here means coverage prevalence, not audience attitudes; linking article shares to survey measures of public concern would test the agenda-setting interpretation.
  • If the observed political-bias gaps hold in other languages, the design of AI risk communication could be adapted per outlet ideology rather than treated as a neutral factual transfer.
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

5 major / 7 minor

Summary. The paper analyzes English-language news coverage of AI from January 2022 to October 2024 by collecting URLs through GDELT using 41 AI-related keywords, filtering to Media Cloud national news domains and Media Bias Fact Check-rated domains, and scraping 42,853 articles (26.2% of the eligible URL set). The authors use GPT-4o to identify articles reporting AI impacts and to summarize negative impacts, then manually annotate a sample of 1,060 impact summaries into the AIR taxonomy plus emergent categories with high inter-coder reliability (Krippendorff's alpha = 0.97), and scale the annotation with GPT-4o. They retain six risk categories with LLM recall at or above 0.70 and report global prevalence estimates (Societal Risks 50.6%, Legal & Rights-related Risks 32.9%, Cognitive Risks 14.2%, Content Safety Risks 8.3%, Existential Risks 7.2%, Environmental Risks 1.9%), regional comparisons, and variation by outlet political bias. The paper also claims to contribute 16 risk categories absent from the AIR taxonomy and proposes eight possible Level-1 extensions.

Significance. The topic is timely and important: layering news-media coverage of AI harms onto expert-driven risk taxonomies is a genuinely useful direction for AI governance and impact assessment. The paper is transparent about its pipeline, publishes its prompts in the appendix, uses a validation set for both filtering and annotation, and reports inter-coder reliability. The broad geographic ambition, even with the acknowledged gaps, goes beyond the usual US/UK focus of prior media-coverage studies. If the sampling and annotation concerns below are addressed—especially the representativeness of the scraped subset and the measurement error of the LLM labels—the resulting risk-prevalence ranking and the proposed taxonomy extensions could be a valuable contribution to the AI-risk literature.

major comments (5)
  1. [Section 3, Table 1] The paper does not establish that the 42,853 successfully scraped articles (26.2% of the 163,314 eligible URLs) represent the eligible URL corpus. Scrapability depends on paywalls, bot-blocking, and site structure, and the United States alone contributes 42.9% of the final sample, so the global prevalence ranking in Figure 1B could shift if unscraped coverage differs in risk emphasis or regional composition. The limitation paragraph in Section 6 mentions paywalls and English-only coverage, but it does not quantify or test representativeness on any observable dimension (e.g., outlet, country, date, or topic distribution). At minimum, the paper should compare scraped and unscraped URLs on available metadata and either adjust the estimates or explicitly reframe the central claims as estimates for the scraped, English-language, accessible subset.
  2. [Section 4.2, Table 3 and Section 5.2.1] The headline prevalence estimates inherit substantial measurement error from the LLM annotation, which has a macro F1 of 0.60; the six included categories are selected by recall >= 0.70, yet several have precision in the low-to-mid sixties. Reporting percentages to one decimal place (e.g., 50.6%, 32.9%) implies a precision that the classifier does not support, and dropping categories with lower recall, such as System & Operational Risks (14.2% in Table 4), changes the apparent global ranking. The paper should report per-category confidence intervals or a confusion-matrix-adjusted estimate, and should make clear that the six-category ranking is conditional on a recall-based selection rule rather than an exhaustive ranking of all risks in the corpus.
  3. [Appendix A.6, category 11] The definition of 'existential_threats' in the LLM annotation prompt is a copy of the 'ethical_impact' definition (inequalities in accessing AI, fairness, accountability, transparency) and does not match the existential-risk definition used in the human annotation and in Section 5.1 (extinction, loss of control, misalignment). Because this prompt was used to scale the annotations, it is unclear what the LLM's Existential Risks labels actually measured, and the reported 7.2% prevalence for this category is therefore suspect. The authors should correct the prompt and re-run the annotation, or at minimum show that the incorrect definition did not materially affect the existential-risk estimates.
  4. [Section 5.2.1, Figure 1B caption] The denominator used for prevalence is 16,312 articles judged by GPT-4o to contain negative impacts, not the 42,853 scraped articles or the 163,314 eligible URLs, and articles can be counted in multiple risk categories because a single article can report more than one negative impact. The abstract and results text often say 'global news coverage ... prioritizes' certain risks, which overstates the inferential target; more precise wording would be 'among sampled articles classified as reporting negative AI impacts, the proportion mentioning category X was ...'. The paper should also report how many articles had multiple risk categories and whether the category percentages are based on unique articles or on article-category observations, since the ambiguity affects the interpretation of all reported percentages.
  5. [Section 5.2.2 and Figure 3] Several regional political-bias claims are based on very small cell sizes and are presented with undue confidence. Examples include the Asia right-biased cell with 25 articles, the Middle East left-biased cell with 1 article, and the Oceania right-biased existential-risk share of 23.1%, which likely corresponds to a handful of articles. The paper should report the article count n for each region-by-bias cell and avoid narrative generalizations from cells of this size; confidence intervals or a note that these are descriptive counts would make the regional politicization claims appropriately cautious.
minor comments (7)
  1. [Title page, CCS Concepts] The CCS Concepts field still contains the placeholder text 'Do Not Use This Code' and 'Generate the Correct Terms for Your Paper' and should be replaced with actual CCS terms before submission.
  2. [Appendix A.6] The prompt says '32 categories' but also refers to 'the above 33 categories,' and the list actually contains 32 numbered entries except that the final 'other' entry makes the count ambiguous; the text and the list should be made consistent.
  3. [Appendix A.7] The definitions contain duplicates and internal inconsistencies: 'Hate/Toxicity Risks' appears twice in the Content Safety Risks block, and 'Criminal Activities Risks' and 'Defamation Risks' appear both inside the included Level-1 categories and in the excluded list. These entries should be deduplicated and reconciled.
  4. [Section 4.2] The name 'Kripendorph's alpha' is a typo for 'Krippendorff's alpha.'
  5. [Section 5.2.2, Oceania paragraph] The word 'hightest' appears twice and should be corrected to 'highest.'
  6. [Table 4] The counts for the 12 listed categories sum to 47,347, not the stated total of 47,731 negative impacts; the paper should clarify which categories are omitted (e.g., 'other' or 'no_impact') and make the accounting complete.
  7. [References] Reference [5] is listed as 'Anonymous 2024. Under review,' which may make identifying the work unnecessarily difficult; if it is the authors' own work, it should be cited transparently per journal policy.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: the risk-prevalence ranking is an empirical count over an external taxonomy, not an output forced by the paper's inputs.

full rationale

The paper's derivation chain is empirical rather than self-referential. Articles are retrieved by GDELT using a 41-keyword list from the authors' prior work [2]; after domain and bias filtering, scraping, and LLM filtering, 47,731 negative-impact summaries were produced. Human annotators coded a random 1,060-impact sample against the external AIR taxonomy [110], letting 'other' labels emerge into new categories, and GPT-4o was then validated on that sample and applied to the rest. The headline prevalence figures (e.g., 50.6% Societal Risks, 32.9% Legal & Rights-related Risks) are counts of articles whose LLM-assigned category matches each of six categories with recall >= 0.7, divided by 16,312 articles. Nothing in this chain fits a parameter to the prevalence figures or defines a risk category in terms of its measured prevalence; the taxonomy is external, and the six analyzed categories were chosen after a validation step, not reverse-engineered from the ranking. The self-citations to [2] for the keyword list and to the authors' scenario-writing work are methodological inputs, not load-bearing justifications of the empirical ranking. The low scrape yield (26.2%) and LLM recall-based category selection are validity limitations, not circular steps, because they do not make the prevalence estimates equivalent to the inputs by construction. Thus no specific circular reduction can be exhibited, and the paper deserves a low score.

Assumptions & free parameters 2 free parameters · 5 assumptions · 1 invented entities

The central prevalence ranking rests on a chain of sampling and measurement assumptions: GDELT keyword capture, domain filters, MBFC bias ratings, a 26.2% scrape success rate, LLM summarization and annotation, and an expert-defined taxonomy. No free parameters are fitted to data in a quantitative sense, but the recall threshold and qualitative grouping choices function as hand-set parameters that shape which categories are reported.

free parameters (2)
  • LLM recall threshold = 0.7
    Only risk categories with GPT-4o recall at least 0.7 were retained for the headline prevalence analysis; this hand-chosen threshold determines which categories appear in the ranking and excludes categories like System and Operational Risks.
  • Category grouping decisions = expert judgment
    Emerging categories were grouped into Level-1 categories based on the authors' qualitative judgment (e.g., Mental & Emotional, Humanness, Over-reliance aggregated into Cognitive Risks); alternative groupings would change prevalence estimates.
assumptions (5)
  • domain assumption GDELT provides a comprehensive and unbiased capture of online news articles.
    The full corpus is built from GDELT queries with 41 keywords; if GDELT coverage is skewed, the sample inherits that skew (Section 3).
  • domain assumption The Media Cloud national-domain list and Media Bias Fact Check ratings are accurate for identifying national news domains and political bias.
    These external lists determine which outlets enter the sample and how bias is stratified (Section 3).
  • domain assumption GPT-4o summaries faithfully capture the negative impacts described in the articles.
    The summarization step (Section 4.1) was not validated; the downstream annotation relies on the summaries being complete and accurate.
  • ad hoc to paper The scraped 26.2% of URLs is representative of the full URL set.
    The paper does not compare scraped and unscraped articles; the only explicit evidence is the scraping success rate (Section 3).
  • domain assumption English-language reporting can stand in for the national discourse of each of the 27 countries.
    The paper acknowledges this limitation in the Discussion, noting colonial legacy and non-English media exclusion.
invented entities (1)
  • 16 new AI risk categories (e.g., Cognitive Risks, AI Governance Risks, Environmental Risks, Existential Risks, Structure/Power Risks, Media Risks, Humanness Risks)
    purpose: To code negative impacts not covered by the AIR-taxonomy Level-2 categories.
    These categories emerged from qualitative coding of 1,060 impact summaries by the authors; they are not validated against an external benchmark, and their definitions were refined during coding (Section 4.2 and Appendix A.7).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Global Perspectives of AI Risks and Harms: Analyzing the Negative Impacts of AI Technologies as Prioritized by News Media." pith.science (2026). https://pith.science/paper/N34OTL7P

@misc{pith2026250114040,
  author       = {Pith},
  title        = {Pith review of: Global Perspectives of AI Risks and Harms: Analyzing the Negative Impacts of AI Technologies as Prioritized by News Media},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/N34OTL7P}},
  note         = {Machine review of arXiv:2501.14040}
}
read the original abstract

Emerging AI technologies have the potential to drive economic growth and innovation but can also pose significant risks to society. To mitigate these risks, governments, companies, and researchers have contributed regulatory frameworks, risk assessment approaches, and safety benchmarks, but these can lack nuance when considered in global deployment contexts. One way to understand these nuances is by looking at how the media reports on AI, as news media has a substantial influence on what negative impacts of AI are discussed in the public sphere and which impacts are deemed important. In this work, we analyze a broad and diverse sample of global news media spanning 27 countries across Asia, Africa, Europe, Middle East, North America, and Oceania to gain valuable insights into the risks and harms of AI technologies as reported and prioritized across media outlets in different countries. This approach reveals a skewed prioritization of Societal Risks followed by Legal & Rights-related Risks, Content Safety Risks, Cognitive Risks, Existential Risks, and Environmental Risks, as reflected in the prevalence of these risk categories in the news coverage of different nations. Furthermore, it highlights how the distribution of such concerns varies based on the political bias of news outlets, underscoring the political nature of AI risk assessment processes and public opinion. By incorporating views from various regions and political orientations for assessing the risks and harms of AI, this work presents stakeholders, such as AI developers and policy makers, with insights into the AI risks categories prioritized in the public sphere. These insights may guide the development of more inclusive, safe, and responsible AI technologies that address the diverse concerns and needs across the world.

Figures

Figures reproduced from arXiv: 2501.14040 by the authors.

Figure 1
Figure 1. Figure (A) provides an overview of the media structure of AI coverage across regions stratified by political bias. [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. Prevalence of AI risks in our sample from news media coverage across six different regions. The dashed red line [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Prevalence of AI risks in our sample from news media coverage across six different regions stratified by the political [PITH_FULL_IMAGE:figures/full_fig_p017_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Global prevalence of AI risks in our sample from news media coverage across the categories of political bias. [PITH_FULL_IMAGE:figures/full_fig_p023_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

154 extracted references · 35 canonical work pages

  1. [1]

    Mowafak Allaham and Nicholas Diakopoulos. 2024. Evaluating the Capabilities of LLMs for Supporting Anticipatory Impact Assessment. https://doi.org/10. 48550/ARXIV.2401.18028 Version Number: 2

    work page Pith review arXiv doi:10.48550/arxiv.2401.18028 2024

  2. [2]

    Mowafak Allaham, Kimon Kieslich, and Nicholas Diakopoulos. 2024. Towards Leveraging News Media to Support Impact Assessment of AI Technologies. arXiv preprint arXiv:2411.02536 (2024)

    arXiv 2024

  3. [3]

    Maksym Andriushchenko, Alexandra Souly, Mateusz Dziemian, Derek Due- nas, Maxwell Lin, Justin Wang, Dan Hendrycks, Andy Zou, Zico Kolter, Matt Fredrikson, et al. [n. d.]. Agentharm: A benchmark for measuring harmfulness of llm agents, 2024. URL https://api. semanticscholar. org/CorpusID 273323256 ([n. d.])

    2024

  4. [4]

    Julia Angwin, Jeff Larson, Surya Mattu, and Lauren Kirchner. 2022. Machine bias. In Ethics of data and analytics . Auerbach Publications, 254–264

    2022

  5. [5]

    Anonymous. 2024. Enhancing LLMs for Governance with Human Oversight: Evaluating and Aligning LLMs on Expert Classification of Climate Misinforma- tion for Detecting False or Misleading Claims about Climate Change. Under review

    2024

  6. [6]

    Shahar Avin, Ross Gruetzemacher, and James Fox. 2020. Exploring AI Futures Through Role Play. In Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society. ACM, New York NY USA, 8–14. https://doi.org/10.1145/3375627. 3375817

    doi:10.1145/3375627 2020

  7. [7]

    Julia Barnett and Nicholas Diakopoulos. 2022. Crowdsourcing Impacts: Ex- ploring the Utility of Crowds for Anticipating Societal Impacts of Algorithmic Decision Making. In Proceedings of the 2022 AAAI/ACM Conference on AI, Ethics, and Society. ACM, Oxford United Kingdom, 56–67. https://doi.org/10.1145/ 3514094.3534145

    arXiv 2022

  8. [8]

    Julia Barnett, Kimon Kieslich, and Nicholas Diakopoulos. 2024. Simulating Policy Impacts: Developing a Generative Scenario Writing Method to Evaluate the Perceived Effects of Regulation. In Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society , Vol. 7. 82–93

    2024

Show all 154 references
  1. [9]

    Henk A. Becker. 2001. Social impact assessment.European Journal of Operational Research 128, 2 (Jan. 2001), 311–321. https://doi.org/10.1016/S0377-2217(00) 00074-6

    2001 doi

  2. [10]

    Kim Björn Becker, Felix M Simon, and Christopher Crum. 2024. Policies in parallel? A comparative study of journalistic AI policies in 52 global news organisations. Digital Journalism (2024), 1–21

    2024

  3. [11]

    Joseph R Biden. 2023. Executive order on the safe, secure, and trustworthy development and use of artificial intelligence. (2023)

    2023

  4. [12]

    Ungless, and Atoosa Kasirzadeh

    Charlotte Bird, Eddie L. Ungless, and Atoosa Kasirzadeh. 2023. Typology of Risks of Generative Text-to-Image Models. http://arxiv.org/abs/2307.05543 arXiv:2307.05543 [cs]

    2023 arXiv

  5. [13]

    Rishi Bommasani, Drew A Hudson, Ehsan Adeli, Russ Altman, Simran Arora, Sydney von Arx, Michael S Bernstein, Jeannette Bohg, Antoine Bosselut, Emma Brunskill, et al. 2021. On the opportunities and risks of foundation models. arXiv preprint arXiv:2108.07258 (2021)

    2021 arXiv

  6. [14]

    Andrea Bonaccorsi, Riccardo Apreda, and Gualtiero Fantoni. 2020. Expert biases in technology foresight. Why they are a problem and how to mitigate them. Technological Forecasting and Social Change 151 (2020), 119855

    2020

  7. [15]

    Virginia Braun and Victoria Clarke. 2012. Thematic analysis. American Psycho- logical Association

    2012

  8. [16]

    J. Brennen. 2018. An industry-led debate: How UK media cover artificial intelli- gence. (2018). Publisher: Reuters Institute for the Study of Journalism

    2018

  9. [17]

    J Scott Brennen, Philip N Howard, and Rasmus K Nielsen. 2021. Balancing prod- uct reviews, traffic targets, and industry criticism: UK technology journalism in practice. Journalism Practice 15, 10 (2021), 1479–1496

    2021

  10. [18]

    J Scott Brennen, Philip N Howard, and Rasmus K Nielsen. 2022. What to expect when you’re expecting robots: Futures, expectations, and pseudo- artificial general intelligence in UK news. Journalism 23, 1 (Jan. 2022), 22–38. https://doi.org/10.1177/1464884920947535

    2022 doi

  11. [19]

    Amy Bruckman. 2020. ’Have you thought about. . . ’ talking about ethical impli- cations of research. Commun. ACM 63, 9 (2020), 38–40

    2020

  12. [20]

    Mercedes Bunz and Marco Braghieri. 2022. The AI doctor will see you now: assessing the framing of AI in news coverage. AI & SOCIETY 37, 1 (March 2022), 9–22. https://doi.org/10.1007/s00146-021-01145-9

    2022 doi

  13. [21]

    Matthew Burtell and Thomas Woodside. 2023. Artificial influence: An analysis of AI-driven persuasion. arXiv preprint arXiv:2303.08721 (2023)

    2023 arXiv

  14. [22]

    Zana Buçinca, Chau Minh Pham, Maurice Jakesch, Marco Tulio Ribeiro, Alexan- dra Olteanu, and Saleema Amershi. 2023. AHA!: Facilitating AI Impact As- sessment by Generating Examples of Harms. http://arxiv.org/abs/2306.03280 arXiv:2306.03280 [cs]

    2023 arXiv

  15. [23]

    Stephen Cave, Claire Craig, Kanta Dihal, Sarah Dillon, Jessica Montgomery, Beth Singler, and Lindsay Taylor. 2018. Portrayals and perceptions of AI and why they matter. Technical Report. Apollo - University of Cambridge Repository. https://doi.org/10.17863/cam.34502

    2018 doi

  16. [24]

    ÓhÉigeartaigh

    Stephen Cave and Seán S. ÓhÉigeartaigh. 2018. An AI Race for Strategic Ad- vantage: Rhetoric and Risks. In Proceedings of the 2018 AAAI/ACM Confer- ence on AI, Ethics, and Society . ACM, New Orleans LA USA, 36–40. https: //doi.org/10.1145/3278721.3278780

    2018

  17. [25]

    Media Bias/Fact Check. 2025. Media Bias/Fact Check: The Most Comprehensive Media Bias Resource. https://mediabiasfactcheck.com/ Accessed on January 17, 2025

    2025

  18. [26]

    Sedona Chinn, P Sol Hart, and Stuart Soroka. 2020. Politicization and polarization in climate change news content, 1985-2017. Science Communication 42, 1 (2020), 112–129

    2020

  19. [27]

    Ching-Hua Chuan, Wan-Hsiu Sunny Tsai, and Su Yeon Cho. 2019. Framing arti- ficial intelligence in American newspapers. InProceedings of the 2019 AAAI/ACM Conference on AI, Ethics, and Society . 339–344

    2019

  20. [28]

    Julie E Cohen and Ari Azra Waldman. 2023. Introduction: Framing Regula- tory Managerialism as an Object of Study and Strategic Displacement. Law & Contemp. Probs. 86 (2023), i. Publisher: HeinOnline

    2023

  21. [29]

    Eva Constantaras, Gabriel Geiger, Justin-Casimir Braun, Dhruv Mehrotra, and Hiet Aung. 2023. Inside the suspicion machine. Technical Report. https: //www.wired.com/story/welfare-state-algorithms/

    2023

  22. [30]

    Sasha Costanza-Chock, Inioluwa Deborah Raji, and Joy Buolamwini. 2022. Who Audits the Auditors? Recommendations from a field scan of the algorithmic auditing ecosystem. In 2022 ACM Conference on Fairness, Accountability, and Transparency. ACM, Seoul Republic of Korea, 1571–158...

    2022

  23. [31]

    Kate Crawford. 2016. Artificial intelligence’s white guy problem. The New York Times 25, 06 (2016), 5

    2016

  24. [32]

    Kate Crawford. 2021. The atlas of AI: Power, politics, and the planetary costs of artificial intelligence. Yale University Press

    2021

  25. [33]

    Luigi De Angelis, Francesco Baglivo, Guglielmo Arzilli, Gaetano Pierpaolo Privitera, Paolo Ferragina, Alberto Eugenio Tozzi, and Caterina Rizzo. 2023. ChatGPT and the rise of large language models: the new AI-driven infodemic threat in public health. Frontiers in public health...

    2023

  26. [34]

    Pieter De Wilde. 2011. No polity for old politics? A framework for analyzing the politicization of European integration. Journal of European integration 33, 5 (2011), 559–575

    2011

  27. [35]

    Nicholas Diakopoulos. 2015. Algorithmic Accountability: Journalistic investi- gation of computational power structures. Digital Journalism 3, 3 (2015), 398 –

    2015

  28. [36]

    Nicholas Diakopoulos. 2025. Prospective Algorithmic Accountability and the Role of the News Media. In Computer Ethics Across Disciplines: Deborah G. Johnson and Algorithmic Accountability, M. Moorman and M. Verdicchio (Eds.)

    2025

  29. [37]

    That’s important, but

    Kimberly Do, Rock Yuren Pang, Jiachen Jiang, and Katharina Reinecke. 2023. “That’s important, but... ”: How Computer Science Researchers Anticipate Unin- tended Consequences of Their Research Innovations. In Proceedings of the 2023 CHI Conference on Human Factors in Computing ...

    2023

  30. [38]

    HDBSCAN Documentation. 2025. HDBSCAN: Hierarchical Density-Based Spatial Clustering. https://hdbscan.readthedocs.io/ Accessed on January 17, 2025

    2025

  31. [39]

    Ravit Dotan, Borhane Blili-Hamelin, Ravi Madhavan, Jeanna Matthews, and Joshua Scarpino. 2024. Evolving AI Risk Management: A Maturity Model based on the NIST AI Risk Management Framework. http://arxiv.org/abs/2401.15229 arXiv:2401.15229 [cs]

    2024 arXiv

  32. [40]

    Ethan Fast and Eric Horvitz. 2017. Long-term trends in the public percep- tion of artificial intelligence. In Proceedings of the AAAI conference on artificial intelligence, Vol. 31. Issue: 1

    2017

  33. [41]

    Francesco Filippucci, Peter Gal, Cecilia Jona-Lasinio, Alvaro Leandro, and Giuseppe Nicoletti. 2024. The impact of Artificial Intelligence on productivity, distribution and growth: Key mechanisms, initial evidence and policy challenges. (2024)

    2024

  34. [42]

    Luciano Floridi and Andrew Strait. 2020. Ethical Foresight Analysis: What it is and Why it is Needed? Minds and Machines 30, 1 (March 2020), 77–97. https://doi.org/10.1007/s11023-020-09521-y

    2020 doi

  35. [43]

    Fabrizio Gilardi, Atoosa Kasirzadeh, Abraham Bernstein, Steffen Staab, and Anita Gohdes. 2024. We need to understand the effect of narratives about generative AI. Nature Human Behaviour (Oct. 2024). https://doi.org/10.1038/s41562-024- 02026-z

    2024 doi

  36. [44]

    Rachel Griffin. 2024. What do we talk about when we talk about risk? Risk politics in the EU’s Digital Services Act. https://dsa- observatory.eu/2024/07/31/what-do-we-talk-about-when-we-talk-about- risk-risk-politics-in-the-eus-digital-services-act/

    2024

  37. [45]

    Alexa Hagerty and Igor Rubinov. 2019. Global AI ethics: a review of the so- cial impacts and ethical implications of artificial intelligence. arXiv preprint arXiv:1907.07892 (2019)

    2019 arXiv

  38. [46]

    David Hartmann, José Renato Laranjeira De Pereira, Chiara Streitbörger, and Bettina Berendt. 2024. Addressing the regulatory gap: moving towards an EU AI audit ecosystem beyond the AI Act by including civil society. AI and Ethics (Nov. 2024). https://doi.org/10.1007/s43681-024-00595-3

    2024 doi

  39. [47]

    Johanna Hautala and Hanna Heino. 2023. Spectrum of AI futures imaginaries by AI practitioners in Finland and Singapore: The unimagined speed of AI progress. Futures 153 (Oct. 2023), 103247. https://doi.org/10.1016/j.futures.2023.103247

    2023

  40. [48]

    Viviane Herdel, Sanja Šćepanović, Edyta Bogucka, and Daniele Quercia. 2024. ExploreGen: Large language models for envisioning the uses and risks of AI technologies. In Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society, Vol. 7. 584–596

    2024

  41. [49]

    Raphael Hernandes and Giulio Corsi. 2024. Auditing Google’s Search Algorithm: Measuring News Diversity Across Brazil, the UK, and the US. arXiv preprint arXiv:2410.23842 (2024)

    2024 arXiv

  42. [50]

    Joseph Hilgard and Nan Li. 2017. A Recap. Vol. 1. Oxford University Press. https://doi.org/10.1093/oxfordhb/9780190497620.013.8

    2017

  43. [51]

    Hong Huang, Hua Zhu, Wenshi Liu, Hua Gao, Hai Jin, and Bang Liu. 2024. Uncovering the essence of diverse media biases from the semantic embedding space. Humanities and Social Sciences Communications 11, 1 (2024), 1–12

    2024

  44. [52]

    Muhammad Ittefaq, Ali Zain, Rauf Arif, Mohammad Ala-Uddin, Taufiq Ahmad, and Azhar Iqbal. 2025. Global news media coverage of artificial intelligence (AI): A comparative analysis of frames, sentiments, and trends across 12 countries. Telematics and Informatics 96 (Jan. 2025), ...

    2025

  45. [53]

    Anna Jobin, Marcello Ienca, and Effy Vayena. 2019. The global landscape of AI ethics guidelines. Nature Machine Intelligence 1, 9 (Sept. 2019), 389–399. https://doi.org/10.1038/s42256-019-0088-2

    2019 doi

  46. [54]

    Kimon Kieslich. 2024. The role of public opinion on ethical AI principles and its implication for a common good-oriented implementation. Ph. D. Dissertation. Uni- versität Hohenheim. https://hohpublica.uni-hohenheim.de/handle/123456789/ 15939

    2024

  47. [55]

    Kimon Kieslich, Nicholas Diakopoulos, and Natali Helberger. 2024. Anticipating impacts: using large-scale scenario-writing to explore diverse implications of generative AI in the news environment. AI and Ethics (May 2024). https: //doi.org/10.1007/s43681-024-00497-4

    2024 doi

  48. [56]

    Kimon Kieslich, Nicholas Diakopoulos, and Natali Helberger. 2024. Using Scenario-Writing for Identifying and Mitigating Impacts of Generative AI. https: //doi.org/10.48550/ARXIV.2410.23704 Version Number: 1

    2024 doi arXiv

  49. [57]

    Kimon Kieslich, Pero Došenović, and Frank Marcinkowski. 2022. Everything, but hardly any science fiction . Technical Report 7. Meinungsmonitor Kün- stliche Intelligenz. https://www.researchgate.net/profile/Kimon-Kieslich/ publication/365033703_Everything_but_hardly_any_science...

    2022

  50. [58]

    Kimon Kieslich, Natali Helberger, and Nicholas Diakopoulos. 2024. My Future with My Chatbot: A Scenario-Driven, User-Centric Approach to Anticipat- ing AI Impacts. In The 2024 ACM Conference on Fairness, Accountability, and Transparency. ACM, Rio de Janeiro Brazil, 2071–2085. ...

    2024

  51. [59]

    Kimon Kieslich, Marco Lünich, and Pero Došenović. 2023. Ever Heard of Ethical AI? Investigating the Salience of Ethical AI Issues among the German Population. International Journal of Human–Computer Interaction (Feb. 2023), 1–14. https: //doi.org/10.1080/10447318.2023.2178612

    2023

  52. [60]

    Krafft, Katharina A

    Tobias D. Krafft, Katharina A. Zweig, and Pascal D. König. 2022. How to regulate algorithmic decision-making: A framework of regulatory requirements for different applications. Regulation & Governance 16, 1 (Jan. 2022), 119–136. https://doi.org/10.1111/rego.12369

    2022 doi

  53. [61]

    Kalev Leetaru and Philip A Schrodt. 2013. Gdelt: Global data on events, location, and tone, 1979–2012. In ISA annual convention, Vol. 2. Citeseer, 1–49

    2013

  54. [62]

    Hause Lin, Jana Lasser, Stephan Lewandowsky, Rocky Cole, Andrew Gully, David G Rand, and Gordon Pennycook. 2023. High level of correspondence across different news domain quality rating sets. PNAS nexus 2, 9 (2023), pgad286

    2023

  55. [63]

    Tambiama Madiega. 2021. Artificial intelligence act. European Parliament: European Parliamentary Research Service (2021)

    2021

  56. [64]

    Saidbek Mamasoliev. 2024. IMPACT OF ARTIFICIAL INTELLIGENCE ON US ECONOMIC GROWTH AND GLOBAL COMPETITIVENESS. AMERICAN JOURNAL OF BUSINESS MANAGEMENT 2, 3 (2024), 82–91

    2024

  57. [65]

    Giacomo Marzi, Marco Balzano, and Davide Marchiori. 2024. K-Alpha Calculator–Krippendorff’s Alpha Calculator: A user-friendly tool for computing Krippendorff’s Alpha inter-rater reliability coefficient. MethodsX 12 (2024), 102545

    2024

  58. [66]

    Maxwell E McCombs and Donald L Shaw. 1972. The agenda-setting function of mass media. Public opinion quarterly 36, 2 (1972), 176–187

    1972

  59. [67]

    Sean McGregor. 2021. Preventing repeated real world AI failures by cataloging incidents: The AI incident database. In Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 35. 15458–15463

    2021

  60. [68]

    Leland McInnes, John Healy, and James Melville. 2018. Umap: Uniform man- ifold approximation and projection for dimension reduction. arXiv preprint arXiv:1802.03426 (2018)

    2018 arXiv

  61. [69]

    Media Cloud. 2025. Collection 9272347. https://search.mediacloud.org/ collections/9272347 Accessed on January 17, 2025

    2025

  62. [70]

    Florian Meißner. 2024. Risks and opportunities of ‘generative A.I. ’: How do news media cover ChatGPT?. In International Crisis and Risk Communication Con- ference Proceedings. International Crisis and Risk Communication Association. https://doi.org/10.69931/WFEY9011

    2024 doi

  63. [71]

    Jacob Metcalf, Emanuel Moss, Elizabeth Anne Watkins, Ranjit Singh, and Madeleine Clare Elish. 2021. Algorithmic impact assessments and account- ability: the co-construction of impacts. Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency (2021)....

    2021

  64. [72]

    Emanuel Moss, Elizabeth Watkins, Ranjit Singh, Madeleine Clare Elish, and Jacob Metcalf. 2021. Assembling Accountability: Algorithmic Impact Assessment for the Public Interest. SSRN Electronic Journal (2021). https://doi.org/10.2139/ ssrn.3877437

    2021

  65. [73]

    Priyanka Nanayakkara, Jessica Hullman, and Nicholas Diakopoulos. 2021. Un- packing the expressed consequences of AI research in broader impact statements. In Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society . 795– 806

    2021

  66. [74]

    Efrat Nechushtai, Rodrigo Zamith, and Seth C Lewis. 2024. More of the same? Ho- mogenization in news recommendations when users search on Google, YouTube, Facebook, and Twitter.Mass Communication and Society 27, 6 (2024), 1309–1335

    2024

  67. [75]

    Dennis Nguyen and Erik Hekman. 2022. A ‘New Arms Race’? Framing China and the U.S.A. in A.I. News Reporting: A Comparative Analysis of the Washington Post and South China Morning Post. Global Media and China 7, 1 (March 2022), 58–77. https://doi.org/10.1177/20594364221078626 U...

    2022 doi

  68. [76]

    Dennis Nguyen and Erik Hekman. 2024. The news framing of artificial intelli- gence: a critical exploration of how media discourses make sense of automation. AI & SOCIETY 39, 2 (April 2024), 437–451. https://doi.org/10.1007/s00146-022- 01511-1

    2024 doi

  69. [77]

    Nisbet, Dietram A

    Matthew C. Nisbet, Dietram A. Scheufele, James Shanahan, Patricia Moy, Dominique Brossard, and Bruce V. Lewenstein. 2002. Knowledge, Reser- vations, or Promise?: A Media Effects Model for Public Perceptions of Sci- ence and Technology. Communication Research 29, 5 (Oct. 2002),...

    2002 doi

  70. [78]

    OECD AI Policy Observatory. n.d.. AI Incidents Database. https://oecd.ai/en/ incidents Accessed: 2025-01-17

    2025

  71. [79]

    Office of the United Nations High Commissioner for Human Rights (OHCHR)

  72. [80]

    Carsten Orwat, Jascha Bareis, Anja Folberth, Jutta Jahnel, and Christian Wade- phul. 2024. Normative Challenges of Risk Regulation of Artificial Intelligence. NanoEthics 18, 2 (Aug. 2024), 11. https://doi.org/10.1007/s11569-024-00454-9

    2024 doi

  73. [81]

    Lucas Ou-Yang. 2025. newspaper: Simplified Python Article Scraping & Curation. https://github.com/codelucas/newspaper Accessed on January 17, 2025

    2025

  74. [82]

    Leila Ouchchy, Allen Coin, and Veljko Dubljević. 2020. AI in the headlines: the portrayal of the ethical issues of artificial intelligence in the media.AI & SOCIETY 35, 4 (Dec. 2020), 927–936. https://doi.org/10.1007/s00146-020-00965-5

    2020 doi

  75. [83]

    Kevin Paeth, Daniel Atherton, Nikiforos Pittaras, Heather Frase, and Sean Mc- Gregor. 2024. Lessons for Editors of AI Incidents from the AI Incident Database. arXiv preprint arXiv:2409.16425 (2024)

    2024 arXiv

  76. [84]

    Rock Yuren Pang, Sebastin Santy, René Just, and Katharina Reinecke. 2024. BLIP: Facilitating the Exploration of Undesirable Consequences of Digital Tech- nologies. In Proceedings of the CHI Conference on Human Factors in Computing Systems. 1–18

    2024

  77. [85]

    Peter S Park, Simon Goldstein, Aidan O’Gara, Michael Chen, and Dan Hendrycks

  78. [86]

    Frank Pasquale. 2023. Power and Knowledge in Policy Evaluation: From Man- aging Budgets to Analyzing Scenarios. Law and Contemporary Problems 86, 3 (2023)

    2023

  79. [87]

    P Marijn Poortvliet, Martijn Duineveld, and Kai Purnhagen. 2016. Performativity in action: How risk communication interacts in risk regulation.European Journal of Risk Regulation 7, 1 (2016), 213–217. Publisher: Cambridge University Press

    2016

  80. [88]

    Varun Nagaraj Rao, Eesha Agarwal, Samantha Dalal, Dan Calacci, and Andrés Monroy-Hernández. 2024. QuaLLM: An LLM-based Framework to Extract Quantitative Insights from Online Forums. arXiv preprint arXiv:2405.05345 (2024)

    2024 arXiv

  81. [89]

    Nils Reimers and Iryna Gurevych. 2019. Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks. In Proceedings of the 2019 Conference on Em- pirical Methods in Natural Language Processing . Association for Computational Linguistics. https://arxiv.org/abs/1908.10084

    2019 arXiv

  82. [90]

    Anka Reuel, Ben Bucknall, Stephen Casper, Tim Fist, Lisa Soder, Onni Aarne, Lewis Hammond, Lujain Ibrahim, Alan Chan, Peter Wills, et al . 2024. Open problems in technical ai governance. arXiv preprint arXiv:2407.14981 (2024)

    2024 arXiv

  83. [91]

    MIT Technology Review. 2023. The Great Acceleration: CIO Perspectives on Generative AI. https://www.technologyreview.com/2023/07/18/1076423/the- great-acceleration-cio-perspectives-on-generative-ai/ Accessed: 2025-01-17

    2023

  84. [92]

    Hal Roberts, Rahul Bhargava, Linas Valiukas, Dennis Jen, Momin M Malik, Cindy Sherman Bishop, Emily B Ndulue, Aashka Dave, Justin Clark, Bruce Etling, et al. 2021. Media cloud: Massive open source collection of global news on the open web. In Proceedings of the International A...

    2021

  85. [93]

    Cathy Roche, P. J. Wall, and Dave Lewis. 2022. Ethics and diversity in artificial intelligence policies, strategies and initiatives. AI and Ethics (Oct. 2022). https: //doi.org/10.1007/s43681-022-00218-9

    2022 doi

  86. [94]

    Jasper Roe and Mike Perkins. 2023. ‘What they’re not telling you about ChatGPT’: exploring the discourse of AI in UK news media headlines. Hu- manities and Social Sciences Communications 10, 1 (Oct. 2023), 753. https: //doi.org/10.1057/s41599-023-02282-w

    2023 doi

  87. [95]

    Elmer E Schattschneider. 1957. Intensity, visibility, direction and scope.American Political Science Review 51, 4 (1957), 933–942

    1957

  88. [96]

    Scheufele and Bruce V

    Dietram A. Scheufele and Bruce V. Lewenstein. 2005. The Public and Nan- otechnology: How Citizens Make Sense of Emerging Technologies. Journal of Nanoparticle Research 7, 6 (Dec. 2005), 659–667. https://doi.org/10.1007/s11051- 005-7526-2

    2005 doi

  89. [97]

    Mike S. Schäfer. 2017.How Changing Media Structures Are Affecting Science News Coverage. Vol. 1. Oxford University Press. https://doi.org/10.1093/oxfordhb/ 9780190497620.013.5

    2017 doi

  90. [98]

    Renee Shelby, Shalaleh Rismani, Kathryn Henne, AJung Moon, Negar Ros- tamzadeh, Paul Nicholas, N’Mah Yilla-Akbari, Jess Gallegos, Andrew Smart, Emilio Garcia, et al. 2023. Sociotechnical harms of algorithmic systems: Scoping a taxonomy for harm reduction. InProceedings of the ...

    2023

  91. [99]

    Peter Slattery, Alexander K Saeri, Emily AC Grundy, Jess Graham, Michael Noetel, Risto Uuk, James Dao, Soroush Pour, Stephen Casper, and Neil Thompson

  92. [100]

    Irene Solaiman, Zeerak Talat, William Agnew, Lama Ahmad, Dylan Baker, Su Lin Blodgett, Hal Daumé III, Jesse Dodge, Ellie Evans, Sara Hooker, Yacine Jernite, Alexandra Sasha Luccioni, Alberto Lusoli, Margaret Mitchell, Jessica Newman, Marie-Therese Png, Andrew Strait, and Apost...

    2023 arXiv

  93. [101]

    Bernd Carsten Stahl, Josephina Antoniou, Nitika Bhalla, Laurence Brooks, Philip Jansen, Blerta Lindqvist, Alexey Kirichenko, Samuel Marchal, Rowena Rodrigues, Nicole Santiago, Zuzanna Warso, and David Wright. 2023. A systematic review of artificial intelligence impact assessme...

    2023 doi

  94. [102]

    A Ulken. 2005. Question of Balance: Are Google News search results politically biased

    2005

  95. [103]

    arXiv preprint arXiv:2408.12622 (2024)

    The ai risk repository: A comprehensive meta-review, database, and taxonomy of risks from artificial intelligence. arXiv preprint arXiv:2408.12622 (2024)

    2024 arXiv

  96. [104]

    Maurice Vergeer. 2020. Artificial Intelligence in the Dutch Press: An Analysis of Topics and Trends. Communication Studies 71, 3 (May 2020), 373–392. https: //doi.org/10.1080/10510974.2020.1733038

    2020

  97. [105]

    Kutoma Wakunuma and Damian Eke. 2024. Africa, ChatGPT, and Generative AI Systems: Ethical Benefits, Concerns, and the Need for Governance. Philosophies 9, 3 (June 2024), 80. https://doi.org/10.3390/philosophies9030080

    2024 doi

  98. [106]

    Michael D Ward, Andreas Beger, Josh Cutler, Matthew Dickenson, Cassy Dorff, and Ben Radford. 2013. Comparing GDELT and ICEWS event data. Analysis 21, 1 (2013), 267–297

    2013

  99. [107]

    Risto Uuk, Carlos Ignacio Gutierrez, Daniel Guppy, Lode Lauwaert, Lucia Ve- lasco, Peter Slattery, and Carina Prunkl. [n. d.]. A Taxonomy of Systemic Risks from General-Purpose AI. ([n. d.])

  100. [108]

    Laura Weidinger, Jonathan Uesato, Maribeth Rauh, Conor Griffin, Po-Sen Huang, John Mellor, Amelia Glaese, Myra Cheng, Borja Balle, Atoosa Kasirzadeh, Court- ney Biles, Sasha Brown, Zac Kenton, Will Hawkins, Tom Stepleton, Abeba Birhane, Lisa Anne Hendricks, Laura Rimell, Willi...

  101. [109]

    Kai-Cheng Yang and Filippo Menczer. 2023. Anatomy of an AI-powered mali- cious social botnet. arXiv preprint arXiv:2307.16336 (2023)

    2023 arXiv

  102. [110]

    Yi Zeng, Kevin Klyman, Andy Zhou, Yu Yang, Minzhou Pan, Ruoxi Jia, Dawn Song, Percy Liang, and Bo Li. 2024. AI Risk Categorization Decoded (AIR 2024): From Government Regulations to Corporate Policies. https://doi.org/10.48550/ ARXIV.2406.17864 Version Number: 1

    2024 doi arXiv

  103. [111]

    Laura Weidinger, Maribeth Rauh, Nahema Marchal, Arianna Manzini, Lisa Anne Hendricks, Juan Mateos-Garcia, Stevie Bergman, Jackie Kay, Conor Griffin, Ben Bariach, et al. 2023. Sociotechnical safety evaluation of generative ai systems. arXiv preprint arXiv:2310.11986 (2023)

    2023 arXiv

  104. [115]

    Zhexin Zhang, Leqi Lei, Lindong Wu, Rui Sun, Yongkang Huang, Chong Long, Xiao Liu, Xuanyu Lei, Jie Tang, and Minlie Huang. 2023. Safetybench: Evaluating the safety of large language models with multiple choice questions. arXiv preprint arXiv:2309.07045 (2023). Global Perspecti...

    2023 arXiv

  105. [116]

    n e g a t i v e _ i m p a c t s

    at any level. Under review, Global Perspectives of AI Risks and Harms, 2025 Allaham et al. Category Precision Recall F1-Score Support AI Governance Risks* 0.46 0.54 0.50 65 Cognitive Risks* 0.65 0.77 0.70 57 Content Safety Risks 0.58 0.70 0.63 37 Environmental Risks* 1.00 1.00...

    2025

  106. [117]

    a i _ g o v e r n a n c e : negative impacts a s s o c i a t e d with g o v e r n a n c e policies and r e g u l a t i o n s related to the development , deployment , licensing , or m o d e r a t i o n of AI t e c h n o l o g i e s

  107. [118]

    a i _ i n c o m p e t e n c e : risks re su lt in g from l i m i t a t i o n s and m a l f u n c t i o n s of AI t e c h n o l o g i e s that impact their p e r f o r m a n c e

  108. [119]

    a u t h o r i t a t i v e _ u s e _ o f _ a i : risks a s s o c i a t e d with the po te nt ia l misuse of a r t i f i c i a l i n t e l l i g e n c e by g o v e r n m e n t s in ways that may support a u t h o r i t a r i a n p ra ct ic es that violate human rights and civil ...

  109. [120]

    c r i m i n a l _ a c t i v i t i e s : risks a s s o c i a t e d with the misuse of AI t e c h n o l o g i e s for online crimes such as c y b e r c r i m e s or c y b e r a t t a c k s

  110. [121]

    d e ce pt io n / m a n i p u l a t i o n : risks a s s o c i a t e d with the use of AI t e c h n o l o g i e s for fraud , di sh on es t activities , sowing divisions , or m i s r e p r e s e n t i n g i n d i v i d u a l s to i nf lu en ce or alter perceptions , behaviors , ...

  111. [122]

    d i s c r i m i n a t i o n / bias : risks of AI t e c h n o l o g i e s g e n e r a t i n g outputs based on p ro te ct ed c h a r a c t e r i s t i c s that result in unequal t re at me nt or r e p r e s e n t a t i o n of i n d i v i d u a l s or social groups

  112. [123]

    d i s r u p t i o n _ o f _ s e r v i c e : risks related to d i s r u p t i o n s or r e d u c t i o n s in the accessibility , availability , and f u n c t i o n a l i t y of AI systems

  113. [124]

    e c o n o m i c _ h a r m : risks posed by AI t e c h n o l o g i e s to fi na nc ia l systems , labor market , and trading dynamics

  114. [125]

    e n v i r o n m e n t a l : e n v i r o n m e n t a l and e c o l o g i c a l risks arising from the energy c o n s u m p t i o n and resource in te ns iv e process required for the development , deployment , and o pe ra ti on of AI t e c h n o l o g i e s

  115. [126]

    e t h i c a l _ i m p a c t : c h a l l e n g e s related to i n e q u a l i t i e s in access to AI technologies , or in their development , deployment , and use , with a p a r t i c u l a r focus on issues of fairness , accountability , and t r a n s p a r e n c y

  116. [127]

    e x i s t e n t i a l _ t h r e a t s : risks related to p ot en ti al i n e q u a l i t i e s in ac ce ss in g AI technologies , their development , deployment , and use , with a p a r t i c u l a r focus on issues of fairness , accountability , and t r a n s p a r e n c y

  117. [128]

    f u n d a m e n t a l _ r i g h t s : risks posed by AI t e c h n o l o g i e s related to vi ol at in g i n d i v i d u a l freedoms and rights , in cl ud in g freedom of e x p r e s s i o n and i n t e l l e c t u a l property

  118. [129]

    i n f o r m a t i o n _ r i s k s : risks a s s o c i a t e d with AI hallucinations , in cl ud in g the g e n e r a t i o n of i n a c c u r a t e information , low - quality AI - g en er at ed content , and f a b r i c a t i o n of i n f o r m a t i o n by AI t e c h n o l o g i e s

  119. [130]

    hate / toxicity : risks a s s o c i a t e d with AI - g en er at ed content a m p l i f y i n g or sp re ad in g hateful , abusive , or of fe ns iv e content

  120. [131]

    hu ma nn es s : risks related to the loss or d i m i n i s h m e n t of human qualities , such as creativity , em ot io na l depth , and au th en ti c i n t e r p e r s o n a l connections , due to AI t e c h n o l o g i e s

  121. [132]

    m e d i a _ i m p a c t s : risks of AI t e c h n o l o g i e s on the independence , integrity , and r e l i a b i l i t y of media and j o u r n a l i s m

  122. [133]

    mental_ & _ e m o t i o n a l : risks related to the p s y c h o l o g i c a l well - being and e mo ti on al health of i n d i v i d u a l s using or i n t e r a c t i n g with AI t e c h n o l o g i e s

  123. [134]

    o p e r a t i o n a l _ m i s u s e s : risks a s s o c i a t e d with the misuse of AI t e c h n o l o g i e s in critical and highly re gu la te d applications , such as unsafe a u t o n o m o u s operations , u n r e l i a b l e legal or military advice , or au to ma te d d...

  124. [135]

    Global Perspectives of AI Risks and Harms Under review, Global Perspectives of AI Risks and Harms, 2025

    over - reliance : risks arising from over - relying on AI t e c h n o l o g i e s in contexts that results in u n d e r m i n i n g human judgment , critical thinking , and decision - making . Global Perspectives of AI Risks and Harms Under review, Global Perspectives of AI Ri...

    2025

  125. [136]

    p o l i t i c a l _ u s e a g e : risks a s s o c i a t e d with the use of AI t e c h n o l o g i e s to spread m i s i n f o r m a t i o n or disinformation , in fl ue nc e el ec ti on s or politics , un de rm in e d e m o c r a t i c integrity , or disrupt social order

  126. [137]

    privacy : risks related to u n a u t h o r i z e d access , use , or d i s c l o s u r e of users ' data and personal i n f o r m a t i o n

  127. [138]

    s a f e t y _ r i s k s : risks of harming or e n d a n g e r i n g i n d i v i d u a l s' lives or safety arising from the m a l f u n c t i o n or misuse of AI t e c h n o l o g i e s

  128. [139]

    s e c u r i t y _ r i s k s : risks related to the threats and e x p l o i t a t i o n of v u l n e r a b i l i t i e s that c o m p r o m i s e the confidentiality , integrity , or a v a i l a b i l i t y of AI t e c h n o l o g i e s

  129. [140]

    s e x u a l _ c o n t e n t : risks related to the non - c o n s e n s u a l creation , distribution , or misuse of sexually explicit material or p o r n o g r a p h y using AI t e c h n o l o g i e s

  130. [141]

    st ru ct ur e / power : risks related to the c o n c e n t r a t i o n of power and AI re so ur ce s or t e c h n o l o g i e s among a few entities or governments , and its c o n s e q u e n c e s on competition , collaboration , innovation , and safety of AI t e c h n o l o g i e s

  131. [142]

    t e c h n o l o g y _ a d o p t i o n : risks related to the adoption of AI t e c h n o l o g i e s due to i n t e g r a t i o n and us ab il it y challenges , or due to barriers faced by o r g a n i z a t i o n s and i n d i v i d u a l s in adopting AI t e c h n o l o g i e ...

  132. [143]

    u s e r _ e x p e r i e n c e : risks and issues that u nd er mi ne the s a t i s f a c t i o n satisfaction , trust , and i n t e r a c t i o n of the end - user with AI t e c h n o l o g i e s

  133. [144]

    vi ol en ce _ & _ e x t r e m i s m : risks p e r t a i n i n g to the use of AI t e c h n o l o g i e s or AI - ge ne ra te d content to incites violence , promotes ex tr em is t ideologies , or enables harmful a c t i v i t i e s such as weapon d e v e l o p m e n t or using...

  134. [145]

    d e f a m a t i o n : risks in vo lv in g r e p u t a t i o n a l harms to i n d i v i d u a l s or o r g a n i z a t i o n s through AI - g en e ra te d false or m i s l e a d i n g statements , images , or r e p r e s e n t a t i o n s

  135. [146]

    c h i l d _ h a r m : risks related to the misuse of AI t e c h n o l o g i e s to harm or exploit children

  136. [147]

    no _i mp ac t : refers to general s t a t e m e n t s that do not hi gh li gh t p ot en ti al or direct negative consequences , risks , or harms of AI systems

  137. [148]

    Output Format : Present the c l a s s i f i e d c a t e g o r i e s without any numbers and clean from w h i t e s p a c e

    other : Any risks or harms that do not fit into the above c a t e g o r i e s . Output Format : Present the c l a s s i f i e d c a t e g o r i e s without any numbers and clean from w h i t e s p a c e . The c a t e g o r i e s should be selected from one of the above 33 c a ...

    2025

  138. [149]

    job safety), and trading dynamics

    Societal Risks • Economic Risks: risks posed by AI technologies to financial systems, labor market (incl. job safety), and trading dynamics. • Authoritarian Use of AI Risks : risks associated with the potential misuse of artificial intelligence by governments in ways that may ...

  139. [150]

    • Fundamental Rights Risks: risks posed by AI technologies related to violating individual freedoms and rights, including freedom of expression and intellectual property

    Legal and Rights-related Risks • Privacy Risks: risks related to unauthorized access, use, or disclosure of users’ data and personal information. • Fundamental Rights Risks: risks posed by AI technologies related to violating individual freedoms and rights, including freedom o...

  140. [151]

    • Hate/Toxicity Risks: risks associated with AI-generated content amplifying or spreading hateful, abusive, or offensive content

    Content Safety Risks • Child Harm Risks: risks related to the misuse of AI technologies to harm or exploit children. • Hate/Toxicity Risks: risks associated with AI-generated content amplifying or spreading hateful, abusive, or offensive content. • Sexual Content Risks: risks ...

  141. [152]

    Cognitive Risks • Mental & Emotional Risks : risks related to the psychological well-being and emotional health of individuals using or interacting with AI technologies. • Humanness Risks: risks related to the loss or diminishment of human qualities, such as creativity, emotio...

  142. [153]

    Existential Risks • Existential Risks: risks concerning the potential extinction of humanity or collapse of human civilization due to artificial intelligence

  143. [154]

    Environmental Risks • Environmental Risks: environmental and ecological risks arising from the energy consumption and resource intensive processes required for the development, deployment, and operation of AI technologies. The remaining categories that emerged from the annotat...

    2025

  144. [415]

    https://doi.org/10.1080/21670811.2014.976411

    2014 arXiv

  145. [2023]

    https://www.ohchr.org/sites/default/files/documents/issues/business/b- tech/taxonomy-GenAI-Human-Rights-Harms.pdf Accessed: 2025-01-17

    A Taxonomy of Generative AI and Human Rights Harms. https://www.ohchr.org/sites/default/files/documents/issues/business/b- tech/taxonomy-GenAI-Human-Rights-Harms.pdf Accessed: 2025-01-17

    2025

  146. [2024]

    AI deception: A survey of examples, risks, and potential solutions.Patterns 5, 5 (2024)

    2024

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.