AI Risk Observatory: Analyzing AI Disclosures in Annual Reports for Societal Resilience
Summary:This research investigates how annual reports can provide insights into companies' AI risk and adoption disclosures by processing them at scale with LLMs. Analyzing 9,821 annual reports from 1,362 UK-listed companies (2020-2025), the study reveals that the share of reports mentioning AI risk rose from 2.8% to 41.2%, while AI adoption disclosures increased from 13.8% to 45.2%. However, the quality of AI risk disclosures is generally low, with only 4.3% of reports containing substantive disclosure
Background and Motivation
In recent years, the rapid advancement of artificial intelligence (AI) has significantly impacted society. However, how do companies disclose their AI risks, and are these disclosures substantive? This study aims to assess the trends and quality of AI risk and adoption disclosures in annual reports by analyzing the content related to AI at scale.
Methodology
The research team developed a reproducible two-stage classification pipeline to analyze 9,821 annual reports from 1,362 UK-listed companies (2020-2025). The method was validated against 474 human-annotated passages, demonstrating high recall and moderate label-level agreement.
Key Findings
-
AI Risk Disclosure Trends:
- The share of reports mentioning AI risk increased from 2.8% in 2020 to 41.2% in 2025.
- AI adoption disclosures rose from 13.8% to 45.2%.
- Named vendor mentions were dominated by a small set of major providers, with Microsoft leading the way.
-
Sector and Market Segment Differences:
- AIM market reports disclosed AI risk at much lower rates than Main Market reports.
- Sectors such as Energy and Data Infrastructure lagged behind in AI risk disclosure.
-
Disclosure Quality:
- Despite the increase in AI risk disclosures, most disclosures were not substantive. In 2025, only 4.3% of reports contained AI risk disclosures classified as substantive.
Industry Implications and Recommendations
-
For Companies:
- Improve the substantive nature of AI risk disclosures, moving beyond superficial descriptions.
- Strengthen AI risk identification and management, particularly in critical infrastructure sectors.
-
For Regulators:
- Establish stricter AI disclosure standards to ensure disclosures are substantive and actionable.
- Increase oversight of AI risk management, especially in the Energy and Data Infrastructure sectors.
-
For Researchers:
- Further investigate the impact of AI risk disclosures on different industries and society.
- Develop more advanced AI analysis tools to enhance the automated analysis of annual reports.
Conclusion
This study provides a new data-driven perspective on AI's impact on societal resilience by analyzing AI disclosures in annual reports at scale.
— END —Source: ArXiv AI (cs.AI) (2026-10-05)
Tags: #AI Risk #Annual Reports #Societal Resilience #AI Disclosure #LLM Analysis
Editorial & Fact-Checking Note: This article is compiled from primary research, official release documentation, and source papers by the ZICQ Newsroom pipeline with automated entity verification and human editorial review. If you notice any technical inaccuracy, please submit a correction via our corrections policy or email our editorial desk directly.
Community Comments