Why I’m doing this
Good intentions need evidence.
Advertising now moves through data, platforms, models, and automated workflows. Teams need to know what they intended, what the system was allowed to do, who was responsible, and what actually happened.
The Institute for Advertising Ethics (IAE) Governance Lab develops practical standards to help teams put ethical principles into practice. I want to help close the gap between saying an AI system is responsible and showing how it behaved.
My contribution
Turn principles into practical checks.
Sean Meier and I work on Layer 4: Technical Conformance. In plain language, that means defining what a team should be able to show about the data, content, permissions, and decisions its advertising systems use.
We’re working toward a proof record that stays connected as work passes from one system to the next. My contribution is to make the questions clear, identify the evidence needed to answer them, and help people check that evidence.
Five questions
What should the record show?
Our working draft focuses on five kinds of evidence. Each one answers a simple question. Together, they help a team trace what happened and give someone else a way to check it.
Where did it come from?
Trace the source and who handled it.
What use is allowed?
Show the permissions and limits.
Who is responsible?
Name the people who own the decisions.
What changed?
Record how the data or content changed.
How can someone check it?
Give another reviewer a way to repeat the checks.
How the pieces fit
Connect the rules to what happened.
Layer 3 addresses how AI advertising work is governed. Layer 5 addresses how an organization is held accountable. Our Layer 4 work connects them: can the organization show that its systems followed the rules it approved?
That takes several perspectives. Advertising teams, technologists, legal teams, and leaders need a record they can understand and review. No single tool or score answers every question.
Current status
The requirements are still being developed.
The next questions are practical: where should the proof record begin, what is the minimum evidence it should contain, and who should check it? The draft is a starting point for that discussion.
The AI Advertising Trust Stack is developing work, rather than an adopted IEEE standard or a finished certification. The examples under discussion do not endorse a particular technology or product. This page describes my contribution, rather than speaking on behalf of IAE or IEEE.
My support for Sean’s VRFD initiative explores a related question: how people can establish human identity and content authenticity. The Technical Conformance work asks what evidence any advertising system should be able to provide.