AI adoption and attitudes in risk management (Q2 2025)

Initial results from a joint ASIS–ISRM survey led by Andrew Tollinton, Andrew Sheves and Doug Gray.

Over two weeks in late June and early July 2025, the Institute of Strategic Risk Management (ISRM) and ASIS surveyed risk professionals on AI adoption and attitudes. The survey was led by Andrew Tollinton, Chair of the ISRM AI in Risk Management Special Interest Group and SIRV CEO with Andrew Sheves (Member) and Doug Gray (Member). 73 risk professionals responded, and the intent is to repeat the survey quarterly to track change against the Q2 2025 baseline.

Key findings

• High familiarity
  42.5% report being “very familiar” with AI.

• General tools are mainstream
  77% use general-purpose AI tools such as ChatGPT/Copilot regularly (38.4% daily; 38.4% weekly).

• Role-specific tools lag
  48% use AI tools designed specifically for their role regularly (19.2% daily; 28.8% weekly).

• Bottom-up adoption dominates
  58.9% say AI use is primarily a personal choice (11.0% required by management).

• Training is self-directed
  43.8% rely on free self-led learning; 0% report company-provided training in this sample.

• Job security concerns are mixed
  Approximately 30% express concern and 20.5% are unsure; the remaining ~50% appear unconcerned.

How to interpret this

Method and limitations

This survey reflects a voluntary, self-selecting group and should be read as a snapshot of early adopters. Respondents were predominantly senior (executive 44.4%; senior 36.1%) and highly experienced (86.1% with 15+ years in industry).

Themes from respondent commentary

What practitioners are saying

• Risk leaders want clear use policies, monitoring and enforcement for AI tools.
• Several respondents highlight rushed adoption without governance and limited review of AI outputs.
• Some organisations are restricting AI usage to internal/local tools due to sensitivity and confidentiality concerns.
• A recurring view is “AI as a tool, not a replacement”, with human review for accuracy and hallucinations.

What’s next

Quarterly tracking

This is intended as the first in a repeating quarterly survey series to measure change against the Q2 2025 baseline, broaden the respondent base, and better understand barriers for more sceptical or reluctant users. Next survey planned Q2 2026.

Authors

Andrew Tollinton (Chair, ISRM AI in Risk Management SIG); Andrew Sheves (Member); Doug Gray (Member).

Acknowledgements: Dr David Rubens.