The AI Risk Observatory: What Can We Learn from AI Disclosures in Annual Reports About Societal Resilience?

Bart Jaworski; Andrew Sutton

cs.AIOct 1, 2026

The AI Risk Observatory: What Can We Learn from AI Disclosures in Annual Reports About Societal Resilience?

Authors: Bart Jaworski, Andrew Sutton

Organizations: Independent Researcher.

Abstract

Societal resilience research relies on access to useful and actionable data, which motivates our main research question: Can annual reports, processed at scale with LLMs, provide a useful signal about how companies disclose their response to AI? We test this by applying a reproducible two-stage classification pipeline to 9,821 annual reports from 1,362 UK listed companies (2020-2025, with partial 2026 data). We first validate the method against 474 human-annotated passages, finding high recall and moderate label-level agreement. We then report three empirical patterns: (i) between 2020 and 2025, the share of reports mentioning AI risk rose from 2.8% to 41.2%, while AI adoption disclosure also rose, from 13.8% to 45.2%, and named vendor mentions cluster around a small set of major providers led by Microsoft; (ii) disclosure varies substantially by Critical National Infrastructure sector and market segment: AIM reports disclose AI risk at far lower rates than Main Market reports, and sectors such as Energy and Data Infrastructure lag behind the rest in AI risk disclosure; and (iii) harm disclosures are near-absent (seven reports across the entire corpus). We develop a substantiveness classification to assess the quality of the disclosure and find that most AI risk disclosure is not substantive: in 2025, 41.2% of all reports mention AI as a risk, but only 4.3% contain AI risk disclosure we classify as substantive.

Figures & tables

Item	Value
Companies in universe	1,469
Companies with $\geq$ 1 report	1,362
Processed reports	9,821
Analysis years	2020–2025 (+2026 partial)
Validation passages	474
Validation companies	15

Table 1: Corpus and validation summary.

Stage-1 label	Support	Precision	Recall
adoption	147	0.57	0.97
risk	101	0.82	0.97
vendor	33	0.67	0.91
general/ambiguous	159	0.58	0.96
harm	0	0.00	—
none	165	0.93	0.97

Table 2: Per-label precision and recall of the Stage-1 classifier against the adjudicated validation set (474 passages; support is the number of reference-positive passages per label; labels are non-mutually exclusive). The validation set contains no true harm instance: all 40 harm-tagged passages were false positives.

Figure 1: Annual disclosure rates for AI adoption mentions, AI risk mentions, general/ambiguous AI mentions, and AI vendor mentions, 2020–2026 (2026 is partial). The shaded region shows the difference between the AI adoption-mention share and the AI risk-mention share.

Figure 2: AI risk category prevalence by publication year (% of reports), 2020–2026 (2026 partial). Rows are ordered by 2025 prevalence. The color scale is set on the complete years 2020–2025; 2026p values exceed the scale and are shown for reference. Categories such as Information Integrity and Workforce Impacts emerge as frequently disclosed only from 2024.

Figure 3: 2025 vendor mentions by report count, ordered by frequency. Bars are split by substantiveness; labels give the number of 2025 reports mentioning each vendor and, in parentheses, the share of those reports rated substantive (report-level band, Section 3.3 ). Microsoft, Google, Amazon, and OpenAI lead the named-vendor landscape, alongside a long tail of other named providers.

Figure 4: CNI sector share of 2025 reports with AI mentions and AI risk mentions, sorted by AI risk mention share. Numbers in brackets give the count of 2025 reports in each sector (the denominator of each percentage).

Segment	Reports	Any AI (%)	Adoption (%)	Risk (%)
FTSE 100	223	84.8	79.8	68.6
FTSE 350	606	82.8	61.2	63.0
Main Market	703	77.5	55.6	57.8
AIM	171	36.8	30.4	6.4

Table 3: Market segment AI disclosure rates, 2025 (% of reports). Rows are nested (FTSE 100

\subset

FTSE 350

\subset

Main Market); counts are reports, not companies (the FTSE 100 row covers 223 reports from 85 companies).

Figure 5: 2025 disclosure rates by market segment (FTSE 100 constituents; all Main Market issuers, including the FTSE 100; AIM) across mention types. The drop from Main Market to AIM is steepest for risk disclosure.

Year	AI risk, all (%)	AI risk, subst. (%)	Gap (pp)
2020	2.8	0.3	2.5
2021	4.2	0.6	3.6
2022	4.8	0.9	3.9
2023	9.8	1.2	8.6
2024	30.4	3.3	27.1
2025	41.2	4.3	36.9

Table 4: Share of reports with any AI risk mention vs. with a substantive AI risk mention (% of all reports in each year).

Figure 6: AI risk mention share under three filters: all mentions, explicit-only mentions (attribution signal 3), and substantive-only mentions, 2020–2026 (2026 partial). Restricting to explicit attribution lowers the level (30% vs. 41% of reports in 2025) but leaves the growth trajectory largely intact, whereas restricting to substantive signals flattens the curve almost entirely (4.3% in 2025): most of the growth is in disclosure lacking specific mechanisms, controls, or targets.

Appendix figures & tables9 assets

Supplementary material from the paper’s appendix.

Appendix

Year	Reports	Any AI	Adoption	Risk	Gen./Amb.	Vendor
2020	1,007	19.8	13.8	2.8	12.2	3.1
2021	1,328	27.4	20.9	4.2	17.3	4.2
2022	1,853	28.4	22.5	4.8	17.0	5.7
2023	1,905	36.8	25.8	9.8	24.0	7.5
2024	1,828	55.2	36.8	30.4	44.7	14.6
2025	1,561	65.5	45.2	41.2	56.1	19.2

Appendix

Table 5: Annual disclosure prevalence by mention type. Reports is the number of processed reports per year; all other columns are % of those reports. *2026 is a partial year (339 reports filed to date).

Sector	Reports	Any AI (%)	Risk (%)	$\Delta$ YoY
Government	23	87.0	60.9	+2.5
Communications	29	86.2	51.7	+26.7
Chemicals	24	62.5	50.0	+21.1
Water	17	64.7	47.1	+13.7
Finance	657	74.7	47.0	+10.7
Defence	22	77.3	40.9	+14.0

Appendix

Table 6: CNI sector AI disclosure summary (2025). Reports is the number of 2025 reports tagged with the sector (the denominator of the AI and Risk shares);

\Delta

YoY is the change in the Risk share from 2024, in percentage points (—: no 2024 reports). Company counts by primary sector are given in the text; rates for sectors with fewer than 30 reports are indicative only.

Figure 7: Composition of AI disclosure over time, 2020–2026 (2026 partial). Each AI-mentioning report is assigned to exactly one slice in priority order (any vendor mention; otherwise any risk mention; otherwise any adoption mention; otherwise general/ambiguous only), so unlike Table 5 the slices are mutually exclusive and sum to the AI-mention rate. Reports whose only AI mentions are general/ambiguous remain a small, stable share (5–7% of reports; 6.9% in 2025); the growth comes mainly from reports with risk mentions and reports naming a vendor. The dashed line marks 2025, the last complete year.

Figure 8: 2025 risk category prevalence: all mentions vs. substantive-only mentions (% of reports). Dark segments show the share of reports rated substantive at report level ( Section 3.3 ) among those mentioning the category, overlaid on the light all-mentions bar; the light remainder shows the gap between mention volume and substantive content.

Figure 9: Share of reports mentioning the top named AI vendors, 2020–2026 (2026 partial). OpenAI is essentially absent as a disclosed vendor before 2023 (three reports in 2021, none in 2020 or 2022).

Figure 10: AI risk mention prevalence (% of reports) by CNI sector and publication year, 2020–2026 (2026 partial). Finance and Communications show rapid growth; Energy and Data Infrastructure remain low through 2025 (the 2026p cells rest on very few reports). Blank cells indicate sector-years with no reports.

Figure 11: 2025 CNI sector disclosure by mention type (% of reports). Rows are ordered by the share of reports with any AI mention (see Table 6 ). Some sectors show high general/ambiguous mention rates but low risk or vendor disclosure.

Figure 12: Composition of risk-mentioning reports by substantiveness band, 2020–2026 (2026 partial). Moderate disclosure dominates throughout; the mix has been broadly stable since 2023 even as volume has grown.

Type	Substantive	Moderate	Boilerplate
Adoption ( $n$ =705)	33.8%	55.5%	10.8%
Risk ( $n$ =643)	10.4%	78.5%	11.0%
Vendor ( $n$ =299)	64.5%	24.7%	10.7%

Appendix

Table 7: 2025 disclosure substantiveness by classifier type (% of reports with the corresponding mention type, using the report-level band defined in Section 3.3 ).

AI Governance

Artificial intelligence governance studies how laws, institutions, and technical controls shape AI development, deployment, accountability, and social effects.

See all 386 papers

RiskNet: A large-scale dataset of AI risk incidents from news with alignment and multi-dimensional annotations
Jun 7, 2026Leihan Zhang, Wecheng Ye, Xianlong Ma +5AI Risk Management AI Governance
School of Economics and Management, Beijing University of Posts and Telecommunications, 10 Xitucheng Road, Haidian District, Beijing,2026 100876, China. · Beijing Key Laboratory of Multimodal Data Intelligent Perception and Governance, 10 Xitucheng Road, Haidian District, Beijing, 100876, Jun China.
Risk Reporting for Developers' Internal AI Model Use
Apr 27, 2026Oscar Delaney, Sambhav Maheshwari, Joe O'Brien +2AI Risk Management AI Governance
† Work done while at IAPS
The Eticas AI Risk Taxonomy: Open Infrastructure for Operationalizing AI Audits
Jul 2, 2026Gemma Galdon Clavell, Pablo Accuosto, Usman GoharAI Assurance AI Risk Management

Jun 7, 2026J/K move · Enter open · S save

RiskNet: A large-scale dataset of AI risk incidents from news with alignment and multi-dimensional annotations

Leihan Zhang, Wecheng Ye, Xianlong Ma, Haochuan Liu+4

School of Economics and Management, Beijing University of Posts and Telecommunications, 10 Xitucheng Road, Haidian District, Beijing,2026 100876, China. · Beijing Key Laboratory of Multimodal Data Intelligent Perception and Governance, 10 Xitucheng Road, Haidian District, Beijing, 100876, Jun China.

AI Risk Management AI Governance

Open paper AskPDF arXiv

As artificial intelligence (AI) systems are increasingly deployed across socially consequential domains, reports of AI-related harms and failures have grown in frequency and diversity. Although existing governance frameworks articulate high-level principles for responsible AI, large-scale empirical resources for tracking and analyzing real-world AI risk incidents remain limited. Existing incident collections are often manually curated, relatively small in scale, and insufficient for continuous, data-driven monitoring and downstream computational analysis. To address this need, we present RiskNet, a large-scale dataset of AI risk incidents constructed from large-scale multilingual news sources. RiskNet applies a structured pipeline for AI risk news identification, event-level report screening, incident alignment, and multi-dimensional incident classification. The resulting resource organizes dispersed news reports into incident-centered records and provides benchmark datasets for event classification, incident alignment, and incident-level risk labeling. In its current release, RiskNet covers hundreds of millions of source records and yields a large-scale collection of AI risk-related reports, including aligned incident clusters and annotated benchmark subsets. The dataset is also accessible through an online platform for browsing and exploration. We describe the data sources, processing workflow, taxonomy design, and technical validation of the resource. RiskNet is intended to support downstream research on AI safety, governance, risk analysis, and benchmarking, as well as longitudinal and cross-source analyses of AI-related harms. By providing a structured and reusable empirical resource, RiskNet helps bridge the gap between high-level governance principles and the documented realities of AI risk incidents.

Abstract

AI Governance

Explore similar work

RiskNet: A large-scale dataset of AI risk incidents from news with alignment and multi-dimensional annotations

Risk Reporting for Developers' Internal AI Model Use

The Eticas AI Risk Taxonomy: Open Infrastructure for Operationalizing AI Audits