This week we uncovered an assorted collection of signals that focus on how randomness can be used to
Useful randomness & how we play
This week we uncovered an assorted collection of signals that focus on how randomness can be used to obscure and inspire, and how tech breakthroughs are changing how we play with machines. Save some time to poke around at beautiful devices from Soviet history and to draw lots of very wrong cats.
—Alexis & Matt
1: Adversarial image cloaking
The ethical use of technology can be encouraged or enforced in a number of ways. Legislation can create explicit limits on usage, strong cultural norms can create implicit boundaries, economic boycotts can create economic pressure to curb undesirable behavior, and more. One of our favorite approaches is when we see designers and technologists hacking the technology itself to provide adversarial mechanisms against problematic tech. The SAND Lab at the University of Chicago has implemented such a strategy in Fawkes, a piece of software that allows individuals to limit how their own images might be used to track them.
The problem the researchers identified is that many services and platforms used to share photos are then using those photos to train facial recognition models. We see examples of this with Facebook’s training models, as well as much more wide-ranging (and deeply problematic) services like Clearview AI. What Fawkes does is it “cloaks” images, making tiny, pixel-level changes that are undetectable to the human eye. But “if and when someone tries to use these photos to build a facial recognition model, cloaked images will teach the model a highly distorted version of what makes you look like you. The cloak effect is not easily detectable, and will not cause errors in model training. However, when someone tries to identify you using an unaltered image of you, they will fail.”
While Fawkes currently exists as standalone software (links to download are in the paper below), you could imagine that if Apple or Google decided to integrate this approach into native Android or iOS camera code, the feasibility of facial recognition tracking could be radically diminished at a systemic scale.
2: Virtual interfaces for touching real things
Add this one to the list of rumors circulating about Apple’s in-development AR/VR glasses. A recent patent application suggests that Apple is developing a method for any surface to become a virtual touch interface. While AR manages to blend the virtual and the real in terms of what you see, actually interacting with anything in AR space has been awkward. Most solutions involving either tapping on a screen to interact with a real-world object or using gloves or other peripheral devices to track real-world touch. Apple’s patent proposes using infrared heat sensors to detect the heat transferred to objects via human touch, thereby being able to “see” where a user has touched a surface or object. This would theoretically allow for a much more natural AR interaction, where touching or holding an object in the physical world could trigger digital interactions and responses.
3: Being more human, one scene at a time
We’ve previously mentioned “virtual celebrities,” but in this case technology has gone several steps further with some thought-provoking results.
The New York Times reports on a robot named Erica who will star in a film called “b”. The movie doesn’t yet have a director or other stars attached, due mostly to delays caused by Covid-19 concerns, so in the meantime the robot has been rehearsing. Learning lines is the easy part; learning how and why to deliver those lines with emotion has been the focus of her process. In each session she asks questions of human actors to better understand why something should be said differently. The result is likely to feel somewhat human, but we hope that the performance is something uniquely robot-like, letting the robot show us a different perspective on human behavior.
4: Quite literally, a random walk
Like many New Yorkers, we’ve been on a fairly tight lockdown for over 140 days. While we’ve been able to wander around our neighborhood and have had a few remote outings, there hasn’t been much to draw us out to a new place since there’s so little happening to discover. A new app, with a mechanism for generating random walks, might be the thing that gets more of us outside.
Randonautica launched in February with a very simple premise: tell the app where you are and it will randomly generate a set of coordinates for you to go to and look around. If you choose to, you can upload pictures or descriptions of what you see. And that’s pretty much it.
Like most truly exploratory applications, this simplicity was the foundation for a whole set of interactions, memes, and inside jokes that helped create community among its users, called Randonauts. Sometimes it only takes a small, simple prompt to spur discovery and creativity.
TikTok is on the chopping block. Instagram is pointless in lockdown. The best we can do is a hokey piece of software that takes us somewhere unexpected.
5: The politics of leaf blowers
Thanks to Dan Sinker for bringing this fascinating Twitter thread to our attention. In it, Zack Furness tells the story of an attempted ban on leaf blowers in Los Angeles in the 90s, its effect on the Latin American gardeners in the city, the development of an electric leaf blower in response to the ban, and the artistic embellishment of leaf blowers to explore the dynamics of Mexican and American pop culture. It’s a journey! What we loved about the thread is the way that it exposes how any technology takes on political meaning based on its cultural context. As Zack puts it, “Leaf blowers aren’t inherently political. But, like so many other technologies, they become ‘political’ in contexts shaped by whom they are used, how and why they are used, and what they come to signify about labor, design, aesthetics, and power.”
The stories about protestors using leaf blowers – both in PDX and previously in Hong Kong – keep reminding me of a cool art installation by @RubnOrtizTorres that highlights some other significant ways that leaf blowers become political objects in racialized conflicts. https://t.co/8Fjzx5vp8h https://t.co/EYaVgsV72H
6: Playing with the machines
There’s a lot of news and hype out there around Machine Learning techniques, breakthroughs, and applications, but a lot of it can feel esoteric or difficult to grasp. This guide, compiled by Kate Compton, gives some examples of fun, simple, and often silly tools you can use to explore what machine learning can do. Matt’s personal favorite is Edges2Cats, which lets you make fun and horrifying cats like this:
A selection of interactive deep-learning works that students can play with to creative improvisational experiences.
One Soviet retro-future
Andrew Johnson tipped us off to this set of Soviet modular smart home devices from 1987, as featured in Technical Aesthetics magazine, of which we now want every back issue. While nearly every device we own is now an indistinguishable black rectangle, this reminds us of the many possible (and delightfully weird) design alternatives for the technology that surrounds us.
Did you enjoy this issue?