Josh Lovejoy

Built Google's AI-design playbook, ran ethics at Microsoft — on what hype is doing to trust.

Reflective practitioner-philosopher of AI design. Ethics-forward, focused on the margins.

Josh Lovejoy helped found the field of Human-Centered Machine Learning. He started Google's PAIR initiative, ran design for Microsoft's Ethics & Society team, worked on the UX of probabilistic systems at Amazon, and is now at Autodesk. Across all of it he has been circling one problem: how do you design an honest interface to a system that never guarantees a right answer?

His concern now is more pointed. The industry has spent three years overselling, and the bill is arriving as a trust deficit — people who no longer believe what a product tells them, including when it is right. Overclaiming does not just embarrass the vendor. It degrades the whole category.

He'll be giving the inside view: what it actually looks like when these decisions get made, who makes them, and where the pressure to overclaim comes from. And how you earn trust back from people who have already stopped believing you, while the backlash is still building.

You'll take away: a language for the argument you are already having at work. If you have ever tried to explain why a confident-sounding feature is a bad idea and lost people’s attention, Josh gives you the case from someone who has made it inside four of the largest companies in the world. And a read on what the trust deficit means for anything you ship next, whether or not you helped create it.

LISTENS AND READS — On Artificiality - posts and pods. Our own conversation with him, and still one of the best things on human-centered AI design. And also Josh’s site: Paradox Theory.