Jay Mandel × the IAE conversation

Ideas into
evidence.

Articles, conference reflections, and work in progress on a shared question: how do we make trust in AI advertising something people can check?

From the conference / October 2026

The questions
I’m carrying forward.

Following the IAE summit at Omnicom, I’m connecting the discussion to my writing and the evidence work I’m developing with Sean Meier. These are my reflections and next questions.

  1. Start with what happened.

    Real incidents give us something to investigate: the decision, the missing control, and the evidence of a fix. That is the rigor I want to bring to Trust Gap—looking at whether governance works in practice.

    Explore the IAE’s incident-to-governance approach
  2. Keep permission attached.

    An asset changes hands. A model transforms it. An ad reaches a new audience. The useful question at each step is whether the use stayed within what someone actually allowed—and whether we can show it.

    See the permission-to-behavior example
  3. Make responsibility checkable.

    A record should identify the source, the authorized use, the approver, and what changed. It should also give another person a way to inspect the claim, challenge it, and find who can act.

    Explore the working evidence record

The material behind the discussion

Open the slides.
Inspect the example.

Our presentation prepared for the October 12 summit, plus the working record and its source notes.

Keep following the thread

The connected work.

All my latest stories

Continue the conversation

Bring a real example.

A claim you can’t trace. A permission that got lost. A decision nobody owns. Those are useful places to start.

Talk with me

Start a conversation

How can I help?

Ask a question or share what you’re working through.

AI conversation · Opens SuperMeAsk Jay’s AIExplore Jay’s ideas with his SuperMe AI in a new tab.