Skip to main content
The Developer's Podcast

Design by AI: Why we need to hack the algorithm

38 min episode · 2 min read
·
Jyotja Trevarenas

Episode

38 min

Read time

2 min

Topics

Leadership, Design & UX, Marketing

AI-Generated Summary

Key Takeaways

  • Data Exploration vs Exploitation: Reverse AI algorithms to identify missing perspectives rather than replicate past success patterns. In employment systems, this creates diverse teams with multiple viewpoints instead of monocultures, providing more choices when crises require pivoting direction.
  • Virtuous Tornado Design Process: Design with excluded users first, then expand outward through iterations asking who remains excluded. This stress-tests solutions against catastrophic scenarios and unexpected needs, creating systems that work during crises when people lack capacity to adapt themselves.
  • Outliers Drive Innovation: The 20 percent of needs scattered at the periphery of human requirements represent vital few who experience system cracks first. Designing for their struggles creates flexible systems with room for change when mainstream users face unexpected challenges or crises.
  • Radical Inclusive Codesign: Recruit currently excluded people as critical experts who define problems and design processes alongside professionals. This builds invested communities organically, eliminates marketing needs, and creates natural placekeepers who maintain and adapt designs over time without additional effort.

What It Covers

Professor Jutta Trevarenas explains how AI systems trained on statistical averages exclude outliers and minorities, proposing data exploration algorithms and radical inclusive codesign methods to create adaptive, resilient urban environments and products.

Key Questions Answered

  • Data Exploration vs Exploitation: Reverse AI algorithms to identify missing perspectives rather than replicate past success patterns. In employment systems, this creates diverse teams with multiple viewpoints instead of monocultures, providing more choices when crises require pivoting direction.
  • Virtuous Tornado Design Process: Design with excluded users first, then expand outward through iterations asking who remains excluded. This stress-tests solutions against catastrophic scenarios and unexpected needs, creating systems that work during crises when people lack capacity to adapt themselves.
  • Outliers Drive Innovation: The 20 percent of needs scattered at the periphery of human requirements represent vital few who experience system cracks first. Designing for their struggles creates flexible systems with room for change when mainstream users face unexpected challenges or crises.
  • Radical Inclusive Codesign: Recruit currently excluded people as critical experts who define problems and design processes alongside professionals. This builds invested communities organically, eliminates marketing needs, and creates natural placekeepers who maintain and adapt designs over time without additional effort.

Notable Moment

Automated vehicle engines tested in 2013 decided to proceed through intersections when encountering a wheelchair user moving backwards. After receiving more training data, the systems made the same dangerous decision but with increased confidence levels.

Know someone who'd find this useful?

Episode Transcript

Hello, and welcome to The Developer Podcast, where we talk about how to make cities worth living in, which often has to do with the spaces between the buildings as much as the buildings themselves. My name is Christine Murray, editor in chief of The Developer and director of The Festival of Place. As we increasingly use AI to summarize and generate our work, are we oversimplifying, conforming, and edging out the diversity and complexity that make cities worth living in? Today, I'm speaking to Professor Jyotja Trevarenas, the director of the Inclusive Design Research Centre and professor in the faculty of design at OCAD University in Toronto. Jutta established the Inclusive Design Research Centre in 1993 and proactively works to ensure that society is designed inclusively. She's been looking at AI and has noted how it often offers answers and solutions that are typical, popular, normative, predictable, and statistically average. The more data we give it, the more it determines what an average human does and how they behave. But that's not where innovation and great ideas come from, and arguably that's not where great places come from either. At its worst, in the built environment, that could mean places that harm people who are not considered average. Yutta believes it could be possible to retrain these AI models and transform the way they work so that they create places that are inclusive to everyone. I hope you enjoy this conversation. Let's listen in. Hi, and welcome to the developer podcast. Tell me about what it is that you do at the Inclusive Design Research Centre in Toronto. So I'm the director and founder of the Inclusive Design Research Centre or the IDRC. I founded it back in 1993, which was, almost pre Webb and, specifically to look at how can we use the, new affordances of digital systems to create, environments, systems, products that are more inclusive of human diversity. So moving from a one size fits all version of universal design to the possibility of recognizing that everybody is different and we need different things and that our lives are changing all the time. So how can we create systems, including, cities, buildings, but especially digital products so that they work for that full range of human needs. So you spoke recently at the Festival of Place about AI, and this is something that's being increasingly used to make our work easier, our lives easier in the built environment. I mean, we've got designers who listen to this podcast, people in planning. They're all looking at how they can use this, to kind of streamline or speed up some of their processes. Your talk was about what happens when AI gets it wrong. Can you talk a little bit about, you know, AI, which was, you know, probably less on your radar when you started the design center, you know, more, critical now, and how, you know, this, mindset that AI is built from is, intersects with …

Get the full transcript (5,218 words) + summary by email — free

One-time email with the complete transcript and AI summary of this episode. No account needed.

One email, no spam. We’ll also show you what SignalCast does.

Browse all The Developer's Podcast transcripts →

You just read a 3-minute summary of a 35-minute episode.

Get The Developer's Podcast summarized like this every Monday — plus up to 2 more podcasts, free.

Pick Your Podcasts — Free

Keep Reading

More from The Developer's Podcast

We summarize every new episode. Want them in your inbox?

Similar Episodes

Related episodes from other podcasts

Explore Related Topics

This podcast is featured in Best Tech Podcasts (2026) — ranked and reviewed with AI summaries.

You're clearly into The Developer's Podcast.

Every Monday, we deliver AI summaries of the latest episodes from The Developer's Podcast and 192+ other podcasts. Free for one show.

Start My Monday Digest

No credit card · Unsubscribe anytime