Emerging Technologies and Innovation in Practice
Emerging Technologies and Innovation in Practice
This domain tests whether you can adopt new tools without transferring unquantified risk to the public.
Core rules and concepts
- Digital twins and BIM require contract terms on model fidelity, data ownership, interoperability (for example IFC exchange), and who relies on the model. Model accuracy is not the same as design intent.
- AI and machine learning give probabilistic outputs trained on historical data. Require validation against independent data, documented limitations and known failure modes, and human review before use in any safety-critical decision.
- Automation and robotics introduce new hazard profiles: stored energy, unexpected motion, and human-machine separation distances.
- Cybersecurity is an engineering requirement. Operational technology and control system networks need segmentation, patch management, access control, and incident response, because connected infrastructure is critical infrastructure.
- Additive manufacturing and new materials often outrun the design codes. Certification of material properties and qualification testing are prerequisites, not formalities.
- Competence now includes knowing when a tool's output cannot be trusted. The seal holder is accountable whatever produced the number.
Worked example
A vendor offers a machine-learning model to prioritise pipeline inspections. Ask for validation against the utility's own failure history, precision and recall at the proposed threshold, and a list of failure modes the model cannot detect. Adopt it as a prioritisation aid with documented human review and condition-based inspection as a fallback. Using it as the sole basis for inspection intervals transfers an unquantified risk to the public; using it with oversight is defensible.
Traps
Accepting vendor benchmarks as validation; assuming BIM model precision equals design accuracy; ignoring cyber risk during control system upgrades; adopting unqualified new materials without testing.