AI
The AI Risk Observatory: What Can We Learn from AI Disclosures in Annual Reports About Societal Resilience?
Publication: cs.AI updates on arXiv.org
Author: Bart Jaworski
Source published: 05/10/2026, 04:00:00 UTC
Briefing created: 05/10/2026, 04:12:11 UTC
Based on a feed excerpt; full article content has not been acquired. Source evidence retrieved 05/10/2026, 04:03:24 UTC.
Briefings
This briefing analyzes AI risk disclosures in UK listed company annual reports (2020-2025) using an LLM-based classification pipeline. Key findings show AI risk mention rates rose from 2.8% to 41.2% between 2020-2025, with substantial variation by sector (e.g., AIM vs Main Market) and industry (Energy, Data Infrastructure lagging). A critical finding is that while 41.2% of reports mention AI risks, only 4.3% are classified as substantive disclosures, indicating most mentions are superficial. The study validates its method against 474 human-annotated passages, achieving high recall and moderate agreement.
Key points
- Between 2020 and 2025, the share of annual reports mentioning AI risk increased from 2.8% to 41.2% among 1,362 UK listed companies.
- Substantive AI risk disclosure is rare, with only 4.3% of reports classified as substantive despite 41.2% mentioning AI risks.
Context and limitations
Data covers 2020-2025 with partial 2026 data; validation relies on 474 human-annotated passages; substantiveness classification may vary by sector and market segment.
Generated by Bittrees-hosted qwen35-2b sha256:aaf42c8b7c3c on 05/10/2026, 04:12:11 UTC. Based on publisher-provided feed evidence; may not cover the full original article. Check the source for details.
View archived briefing on IPFS
bafybeieg5zzbgicfin3uk3i7lisxj6m5fgmdubg2ocvqjydwgnevrnerwa