Weeknotes 312 – preventing curling agentic AI parenting

Last week was almost fully dedicated to Dutch Design Week (six trends as seen by Dezeen). On Tuesday with the session on the the projects of the Expertisenetwork Systemic Co-Design where I shared and reflected on the progress of the WijkbotKit as a kit for robotic perspectives. Last Sunday an interview Tomasz and I had with Rijnmond came online, which is both a good representation of our goals (apart maybe from the heading that is more representing the question than our narrative :)

On Wednesday, we had a very good ThingsCon Salon to kick off the project for an exhibition of “Generative Things” at the December TH/NGS conference. Sen did a lovely futuring workshop, and we had a nice turnout. Thanks to Stadslab Eindhoven for the kind hosting.

The Design & AI Symposium inspired me (see Triggered Thought), and I visited more of the exhibitions on Friday. As always, I had too little time to see everything.

The coming week will be an exciting week. Not in a good way, looking as we will be the last week before the US elections, and there is still a significant chance Trump will win. We will have all analyses next week on how much AI was part of it or not. Less than expected, etc. And maybe we will only learn that retrospectively. It is part of society for sure and will be the coming period even more. How it will influence society will depend on who is in office.

Who would have thought that, in the end, it is Elon Musk who dominates the elections and maybe the outcome? For me, it feels like a superhero movie like Spiderman, where the fight is against that one supervillain ruling everyone from a capitalistic agenda. The agenda is power and money, not any usual political or societal ideology (check also the Russian connection). And just like the superhero movies, entertainment value defines success. So the big question is, if there is a superhero in this movie…

And I just see this nice post is stated kind of the same feeling.

Triggered thought

Two keynote speakers opened the Design & AI Symposium on Wednesday morning, organized last week during Dutch Design Week by TU/e and TU Delft. Tobias Revell did a lovely talk about AI and magic entanglement. To summarize, one of the messages is that AI is not magic, and claims of scientific underpinnings are more marketing than based on real research. Companies are suggesting research to sell AI as magic. In his view, money and power are dominating AI. There are not really artistic goals.

After him, Bas van de Poel presented his work at Modem. He is a great designer who has found a form of exploring futures by inviting collaborating designers to publish research reports on certain topics. The output is often also a design to underpin visions.

From personal encounters, I can say I respect both of their work a lot. it might seem to represent two visions, but it is also one story with two branches. One of Modem’s works made me think about how AI also drives a new aesthetic in product design. The Terra project is a speculative device that helps you navigate a city without following a specific route. It lets you wander a city, maybe even get lost, with the security that you will reach your end goal. This sentence already has an interesting concept: Are you free to wander if getting lost is not possible? But that is something to explore another time, maybe.

Terra is not a new idea. It is probably for Modem and their partner Panter&Tourron, as the earlier example was made by a startup in 2015 that failed to execute successfully. The company created a smart Bluetooth device called Wander that was doing the same: a device that looked like a watch, only helping you with a pointer to the end goal, not offering any routes. It was up to you to decide how much you would like to wander in that direction. The device was too complex to make, and the app that was the pivot was not exciting enough, so the startup adventure ended in 2019.

Terra is not just a new iteration; it is building a new product based on the current state of our relationship with technology. It does not leave it all up to you to decide; it offers opportunities to use the capabilities of generative AI to form a partnership in getting lost and wandering without panicking. It is, in that sense, representing the often-mentioned co-performance with AI.

But it is also clear that this relationship is very layered. Is it a form of curling parenting, taking away all frictions? It is tempting to connect it to the focus on Agentic AI. What can we learn from these types of first explorations to design these new agents?

In 2019, when I was diving deep into predictive relations with things that predict (and the potential alienating consequences), I made a first draft of a design approach. See also this short paper.

  • Deconstruct what a smart product makes smart (or intelligent)
  • Think about the impact of adding predictive intelligence to our interactions with that thing. In current times you would say, what would be the impact of a thing getting generative capabilities
  • Design engaging interactions between human and AI based on respectful co-performance.

What I like to focus on and -as note to myself- would be good to revisit, is the first rough design approach. For me, the last part was key to investigate, and IMHO I think this is still key and not a standard part of the routine. Designers often stop at the Magic of step two, not diving into the consequences of step 3. That should also trigger deeper thinking about the consequences and dilemmas of agency.

To end, some wishful thinking. Would the rise of generative AI and its mundane use train us not to overtrust technology? We might have a bumpy learning curve, but let’s hope that we will end with the continuous realization that the answers our AI agents provide might be based on hallucination, and we need to always add some grain of salt before blindly following the opinions of the tools we are offering. And could that even teach us a habit of being more critical in the polarised synthetics society? Let’s hope.

Read the full newsletter here, with

  • Notions from last week’s news on Human-AI partnerships, Robotic performances, Immersive connectedness, and Tech societies
  • Paper for the week
  • Looking forward with events to visit

Buy Me a Coffee at ko-fi.com