Soft Tech, Human Values: How Gehl is Reimagining AI in Urban Planning

This text was written by humans.

At Gehl, we have a strong tradition of practicing value-based and human-centered urban planning. As AI technologies are rapidly breaking into our ways of working, we’ve started asking ourselves: how can we use AI to enhance our understanding of human values, rather than letting our work be dictated by the past narrative synthesized by AI products?

With the rise of commercially available tools, such as ChatGPT, Midjourney, and Gemini (formerly Bard), we’ve seen an increasing interest in these technologies from clients and internally at Gehl. It’s exciting to watch colleagues chat with ChatGPT or create visual references with Midjourney! But as these sparkly, magic AI buttons work their way into our daily workflows, it’s time to pause and reflect upon how we truly feel about these tools.

Fast pace to slow paste: balancing AI innovation with human reflection

As members of Gehl’s R&D team, we are not shy when it comes to diving head-first into testing new technologies. But we also acknowledge that the current AI discourse has been driven by the rapid commercialization of a few models, and is not necessarily representative of artificial intelligence research as a whole. It is well-known that many of the popular models have problems such as plagiarism, bias, and representation issues, and while this might be an inherent value conflict, we also believe it’s the result of the AI ‘space race.’ To address this at Gehl, we are working to introduce moments of reflection and bring our human-centered approach and values into technological discussions.

In the summer of 2023, we held a workshop about artificial intelligence at the Gehl summer gathering. Through analog tools like communal mapping exercises, collaging, and group discussions, we explored our hopes and fears about the future of AI. This allowed us to understand not only the general excitement about the possibilities of AI technologies but also the mysteriousness that seems to obscure them. From depictions of blindfolded men on unicorns to mysterious wizards, Gehlers found common creative ground to explore our emotions and decelerate this fast-paced topic through the slow pace of analog tools.

We synthesized our learnings and reflections from the workshop and produced a small booklet, which playfully describes the AI opportunities and challenges our company faces. We hope these learnings guide ongoing discussions and that our ideas can continue to evolve as we keep being creative together.

Reimagining AI Through Play: The ‘What We Cherish’ Experiment at COP28

Last December, we saw another opportunity to explore AI at kyu House, the collective’s week of programming at COP28 in Dubai. Building on the insights from our internal workshop, we decided to reframe the hard-edges and mystery that can obscure AI by engaging people through play and curiosity. As part of our ‘Softer Futures’ session, we introduced ‘What We Cherish,’ a digital experiment with technology, sound, and storytelling. It utilizes different AI algorithms to interpret the emotional tones and themes expressed through spoken language and presents these results through soft shapes and playful melodies.

We wanted to use AI to enhance our understanding of human values — rather than letting AI dictate them. Instead of following the paths laid by Big Tech, we went down GitHub rabbit holes and scoured online communities and documentation for small, friendly, and open-source solutions, particularly those driven by empathy rather than growth and performance.

‘What We Cherish’ is by no means a finished packaged deal; it’s a continued discussion of what we first uncovered last summer and an experiment we hope to evolve, reiterate, and reconfigure in future works.

Embracing softness: a human-centered approach to AI

Rather than renouncing AI technologies and isolating our practice to long-established tools and methods, we want to embrace softness as a way of reframing how we engage with our machines, algorithms, and codes. We’re drawn to the concept of ‘soft technology,’ which means working towards more inclusive, ever-changing, and participatory interfaces, over the hard edges of efficiency, precision, and automation. This approach is not new, but AI has certainly increased the urgency of these kinds of conversations.

Just as the human and life-centered approach to urban planning and design is core to Gehl, the way we engage with technology should follow suit. Gehl is not a technology company, but we have a strong tradition of value-based design that we aim to extend to all corners of our work.  By adopting this softer approach when working with AI, we have learned to listen to each other in more thoughtful ways and attuned our tools to identify hidden patterns.

Gehl’s 7 principles for working with AI

Looking to apply our approach to your own work? Here are our seven guiding principles when engaging with AI:

01. Be transparent and open

Accredit the tools you use and the extent to which you use them.

02. Be aware of its built-in bias and value systems.

Always question the algorithm.

03. Avoid a binary framework.

Think beyond utopia and dystopia. Explore the protopia.

04. Embrace softness and slowness in tech.

Just like we do in public space.

05. Consider the human and planetary cost of the systems you use.

Think about the data centers and data labelers.

06. Use and contribute to open source when possible.

This helps build and democratize knowledge.

07. Don’t forget analog tools.

New tech will always be introduced, so it’s important to keep being creative together.

This article was originally developed for the kyu network internal blog under the title ‘Sparkle Prompts, Unicorn Cowboys, and Soft Tech.’