NEW The 2026 State of AI Trust Report is now available. Read the report →

Intelligence & Research

See AI risk before it becomes AI headlines

Market, regulatory, and threat intelligence, industry benchmarking, and original research on AI security, governance, and the risks of emerging AI.

Advisor walking a client team through a written assessment.

Flagship publication

The 2026 State of AI Trust Report

Our annual assessment for boards, regulators, and enterprise AI leaders. Fourteen chapters, 25 defining statistics, more than 40 documented case studies, and five original frameworks — synthesised from Stanford HAI, McKinsey, Deloitte, PwC, IBM, Edelman, Gartner, NIST, the EU AI Office, Singapore’s IMDA, the UK AI Security Institute, OWASP, MITRE, the AI Incident Database, and court records.

  • Governance and board oversight
  • Risk taxonomy and heatmap
  • Security, injection, and red teaming
  • Assurance and certification
  • Regulatory divergence
  • Incidents and case studies
Read the report

How we publish

Research standards we hold ourselves to

Method disclosed
Sample, period, and method are stated. A finding without a method is an opinion with a chart.
Limits stated
Where a sample is small, unrepresentative, or self-selected, we say so in the finding rather than the appendix.
No client attribution
Client engagements never appear as case material without written consent, and never in identifiable form without it.
Coordinated disclosure
Security findings affecting identifiable products follow our responsible disclosure policy before publication.
Corrections published
Corrections are issued against the original publication and recorded, not silently edited in.

Deploy AI with confidence.

Start with an independent assessment of your highest-stakes AI system.