Trust and Failure in Technology
One way we’re failing to establish trust in technology.
In my day job, I design videogames that support mathematics learning. Despite my training in more traditional classroom-based mathematics teaching and learning, I chose videogames as a platform specifically because people playing a good game have trust in the game. When you experience failure during gameplay, you look automatically for feedback paired with that failure — because a good game won’t just arbitrarily fail you. There’s a reason for that failure, and the game will tell you what the reason is.
I believe that trust is a key component to all learning experiences, in classrooms, libraries, maker spaces, videogames, etc. We have to be able to trust our products, our colleagues, our professors, our friends, in order to learn from and with them. And trust requires a lack of arbitrary punishment — failure needs to come paired with useful feedback for understanding and correcting that failure. This trust — and the failure paired with feedback — is crucial for designing anything, whether learning environments, videogames, digital products, or any other part of the human experience.
As I wrote about the Design Contract, riffing off the brilliant Brousseau and extending from learning to more general design:
With the Design Contract, the designer promises that the product is usable, reliable, and trustworthy…and that the product can be used with only whatever prior knowledge the user brings. That prior knowledge will be challenged with new problems and novel contexts, but the user has everything they need within themselves and within the product to accomplish their goals, whether playful or productive (or both).
In other words, when a user begins using a product (or a student attends a class, or a gamer logs into Battle.net), that user knows that they can use/learn/play successfully. They know that if they do something wrong, they will be notified about their failure, and that failure will be paired directly with feedback, so they understand why and how the failure occurred (and how to avoid it in the future). This failure and feedback, in a well designed product or context, will never be arbitrary — it will always be trustworthy.
In fact, safety from arbitrary nonsense is just a key component of living a good, contented, reasonable life. When we click a button, we want the same thing to happen every time we click that button— the TV turns on, the email is deleted, your avatar jumps, the microwave adds 30 seconds to the timer. When we come to class every week, we want our professor to give us the same credit for attendance that the other students who come each week get; when we register to vote, we want the same access to the ballot box that a voter in a different neighborhood, a different county, a different state gets; when we live as our best selves, we want the same access to safe appropriate bathrooms and good healthcare that other genders get. We just want to trust that we, as people, learners, voters, and players, get a fair shot at this game called life — I could wax eternally philosophic about this, but I’m only tackling a reasonably sized topic today: trust in technology.
When the pandemic dropped, our lives became more reliant on technology than ever before. As a professor, in order to teach a reasonably good class in this new world, I have to be well versed in a long stack of layered tech — here’s just a couple off the top of my head:
the suite of Google products (our institutional providers of email, cloud storage, etc.)
Canvas (our learning management system)
Zoom for our synchronous class time
Panopto for the storage of our Zoom recordings of our classes for students attending asynchronously
Slack or Discord for asynchronous communication with students
single sign-on through the university to get verified access to all those products above
Duo for multi-factor authentication to verify my single sign-on to get verified access to all those products above
Now, these may not seem like too much — and frankly, I’m good with each of them individually — but the real problems emerge when we have to layer them upon each other.
Here’s a tiny little example: our single sign-on verification expires every day at different times — so I may think I’m logged into my official Zoom, and according to my desktop app I am, but when I arrive in a Zoom meeting I get shoved into a waiting room. This might be because my single sign-on expired — although it’s also possible that I was somehow logged out of my Zoom account, OR I’m accessing the meeting through a personal account by mistake, OR my single sign-on system just isn’t telling Zoom properly that I’m good to go.
(Or I just need to update my operating system. Or turn my computer off and then back on again. Or run outside screaming and go back to my childhood life of peace, quiet, and no electricity.)
In other words, there is distinct failure occurring here — I wanted to trust that if I clicked the right button to join the Zoom meeting, I would, but it arbitrarily failed. And there’s just enough feedback to indicate this arbitrary failure — getting trapped in the waiting room, while the host worries that if they let me in they’ll be subjected to pathetic zoombomb wankers who are just pretending to be me — but no feedback as to why the failure happened.
Was it Zoom? Or my university single sign-on? Or user error? Or lack of communication between Zoom and the single sign-on? Or maybe my Verizon internet hiccuped randomly?
(Or maybe it was sunspots?)
This is a single tiny example, but multiply that by all the products that sometimes are accurately verified and sometimes aren’t, and multiply that by the 120 students in one of my classes all struggling with this every week, and multiply that by weird updates that happen in strange places at strange times. My iPhone — which I use to authenticate via Duo — recently updated the iOS, and now Duo only sometimes recognizes my authentication. Is that the iOS update? A Duo issue? User error? Problems between the phone and Duo? The phone network and Duo? Should I turn my phone off and then on again?
(Should I turn the sun off and then back on again? Maybe that would fix the sunspots?)
As I’ve spent the year thinking about our layers of technology, I’ve realized that failure at one layer in our technology use may be accompanied by feedback, but that feedback just isn’t reaching us, because we’re hanging out at one layer (e.g., the Zoom waiting room), and the failure and feedback is happening at another layer (the single sign-on that expired in some random secret database or whatever), and since the two don’t talk to each other in a well designed fashion with the actual user in mind, I never get the feedback that goes with the failure I’m experiencing. So the failure is absolutely arbitrary to me — and I, like most humans, do not like or trust arbitrary experiences:
If I hit the +30 seconds button on my microwave, and sometimes it works but sometimes it adds 1 second or 30 minutes instead — I’m getting a new microwave.
If I go to class every week and yet my attendance is marked as only coming once a month, I’m going to be scared and angry.
If I hit a button to respond to an email and it deletes the email instead, I’m going to get worried about what emails I may have accidentally deleted, and will have to go dig through my trashed emails to make sure there isn’t something super important in there — and I’m really going to hope that my “respond to this email” button hasn’t accidentally been programmed to sometimes execute permanent deletion.
In fact, I’m going to hope that one of the layers of my tech life didn’t get programmed by a cranky underpaid worker so that when I click “respond to this email” it’s actually auto-sending a nasty email response, muting the original sender, and then permanently deleting the whole conversation. Because that’d be pretty easy to do — the only reason I don’t think that’s happening, is because I have trust in my technology. And I trust that a button that says “respond to this email” actually means what it says. But the more layers that get involved, and the more failure I experience that comes with no feedback, the more I feel that clicking a button will have an arbitrary and undesirable response.
I think of our commitment to these multiple layers of technologies as being a factorial problem. If I use 10 different pieces of technology that connect to each other in some way, then it’s really like using 10! pieces of technology, because we have to account for all the updates, the edges between the technology, the platforms of the various technologies, the fact that my technology wants to talk to your technology, the help pages online that are outdated (or future-dated for a version that hasn’t been rolled out to me yet), the fact that reliable and affordable broadband isn’t considered a necessary service for modern life in my country yet, and that there is very little protection from companies who sell us hardware or software or services that aren’t actually working properly. So buying a new piece of software is not increasing your technology count by 1, but rather increasing from 10! to 11! which is HUGE.
So maybe hold off on that new piece of tech? At least until our technology layers start talking to each other (and us) better — because every new piece adds to our experience of arbitrary, frustrating, unhappy, no-feedback failure.
As I was writing this, the tone changed — it started all playful, because I have these lovely hilarious examples of arbitrary failure, with pretty pictures and everything! Then I ended up having time to write only because an expensive virtual conference I was attending combined three things I hate: 1) an idiosyncratic virtual conference platform that was locked down to make sure that only people who had paid could attend, which resulted in people who DID pay having issues with access; 2) a poor and slow-to-respond tech support system by the platform developers that starts with ‘Did you turn it off and turn it on again?’ and takes hours to escalate to someone who can actually help (I think — we haven’t actually made it to someone who can actually help); and 3) a complete inability to log in, even with official registration, multiple computers/browsers, and a tech support professional who calls me ‘wife’ and in return for that honor, is responsible for fixing my tech! So I wrote most of this in the six hours it took to figure out a secret backdoor to the conference, thinking crankily about the fact that I was missing my team’s first talk, and might have to notify the other 6 sessions I’m involved in that I can’t attend the premier conference in my field. So yeah, it definitely ended not playful — because arbitrary failure without support make life really forking frustrating!
The next piece about tech and trust I write will be goofier and funnier, with all those aforementioned pictures of arbitrary failure, I promise — unless the sunspots start acting up again.
This work is licensed under a CC BY-NC-SA 4.0 license. Attribution must include a link to this work.
