Going Digital, Staying Human: Fast Tips for using AI in impact assessment

The rapid acceleration and ready availability of AI demands that infrastructure professionals upskill quickly. Here, we distill expert advice from I2S' IAIA25 Conference session where we led impact assessment professionals through a rapid-drafting process to share experiences and advice. Here are their fast tips on decision-making critieria helpful for when you are considering whether/when/what/how and with whom to use AI:

AI quality and capacity for proper use
Is there the capacity to correctly and effectively use AI? 

When does it need to be used in our project? Do we have time to develop processes? 

Have we validated our insights/assessments/decisions with affected people? 

Is it trustworthy?

Do communities have the capacity to asses whether AI benefits them?

Do we have technical capacity to create meaningful and reliable analysis from AI? To provide proper oversight? 

Data integrity and security
Do we have good quality data? 

From where has the data been sourced? 

Is the data biased?

Have we considered data/privacy ‘no gos’ – e.g., sacred sites, personal data, identifiable data?
Is the data too sensitive? 

Confidentiality? 

Is data protected? 

Principles/values/ethics
Does the outcome/benefit justify the resources required to apply AI (e.g., power, data servers)? 

Can we demonstrate that AI will ‘make things better’? 

Are there any human rights risks? 

Do we understand potential future risks? 

Does AI or digital technology become a barrier to building relationships (present and future)? 
 

Costs/benefits

Do we know the cost of using AI? The benefit of its impact? 

To what extent can we inform communities/stakeholders/decision-makers about costs? 

Does it benefit humans/humanity? 

Governance and oversight
Have we considered governance arrangements? 

Who are the report-accepting authors (academia, researchers, EPA, government, local level), and are there restrictions on whether they can accept AI-assisted assessments or reports?

Is there strong governance/policy/guidelines in place for use of AI in the assessment? 

Thank you and next steps

Thank you to all of our International Association for Impact Assessment Association for Impact Assessment 2025 Crawford School of Public Policy I2S Workshop participants for sharing your expertise, energy and ideas!

We’re now compiling an IAIA FasTips to support decision-making about AI and digital engagement tool use in impact assessment. Participants also provided their ideas on several other key areas for consideration, in addition to the decision-making criteria/questions noted above, including: 

-Principles we should apply when using (or deciding whether to use) AI in IA
-What questions we should ask ourselves when considering use of AI in assessments
-Who should use AI and who should decide whether AI is used
-Trade-offs that must be considered when using AI in IA.

Missed the session and interested to contribute? Please get in touch!