ACE AI framework: governance and risk takeaways for consultancy engineers
Reviewed by Joe Ashwell

First reported on New Civil Engineer
30 Second Briefing
The Association for Consultancy and Engineering has issued an AI risk framework to guide consultancy firms deploying tools such as generative design, predictive maintenance models and automated code-checking in infrastructure planning and design workflows. The framework focuses on governance, data provenance, model transparency and liability allocation between clients, consultants and software vendors, aiming to structure how firms document assumptions and validate AI outputs against existing standards like Eurocodes. For geotechnical and civil teams, it signals that AI-derived ground models, load assessments and design options will need explicit risk registers and auditable decision trails.
Technical Brief
- Framework is issued by the Association for Consultancy and Engineering (ACE) specifically for consultancy firms in infrastructure.
- Scope explicitly covers AI use in planning and design stages, not construction-phase site operations.
- ACE positions the framework as a risk-management tool rather than a technical design standard.
- Document is intended to sit alongside, not replace, existing safety and quality-management systems.
- Safety focus is on preventing unverified AI outputs entering formal design deliverables or issued drawings.
Our Take
ACE has recently been pushing long-horizon infrastructure thinking through its Delivering Infrastructure 2050 work in Wales and Scotland, so an AI framework from the same body is likely to be framed as a governance tool for very long-lived transport, energy and water projects rather than just back-office efficiency.
With 119 Policy stories and many tagged to Standards/Guidelines and Safety in our database, this is one of the first clearly AI-focused governance pieces in the civil consultancy space, signalling that AI use on projects is starting to be treated like any other safety-critical process that needs codified practice.
Prepared by collating external sources, AI-assisted tools, and Geomechanics.io’s proprietary mining database, then reviewed for technical accuracy & edited by our geotechnical team.


