The click-baity idea I am stewing on here is basically this: do people actually really hate learning about other people?

The inclination I’ve picked up over the years is that business-minded folks seem to loathe the “requirements gathering” and “discovery” phases of research especially, where early ideas are shown to potential users or perhaps highly important users to consider, such as people with disabilities. And there is a general vibe, a bit hard to describe, where the business folks (what you might call the “product owner” whose idea has spawned some new spinoff or product or software or service) seem to really hate involving users. My conjecture is that folks who have some kind of vision often want to be held accountable to nothing, if they could have their way. And if they could have their way, they’d also make things according to their own interests and tastes.

And yet human-computer interaction research, in both industry and academia, hinges on this fundamental gap that people tend to have in their “local” (eg “personal”) knowledge: that other people probably don’t agree with you on everything.

Now, I’m going to talk about two things in HCI research that are both “anti-social” in a way. One is good. The other? Worries me.

HCI research is anti-shithead

So one of the best ways to reign in an absolute spreadsheet of a human, someone who really only cares about whatever little thing they’re on about, is to run a study that demonstrates not only how their core idea is lacking but also ways that their core idea could improve. And to some degree, this is social. That is to say, the HCI research acknowledges that something is better when we learn about more people and perspectives and so on.

But it is also anti-social in a sense, because it fundamentally doesn’t trust some singular person to have a perfect vision and perfect execution. We need to test ideas rigorously. And we do that by involving others. So to the “product owners” at the top, it’s a bit anti-them (or at the very least anti-dictatorship). From the perspective of a terrible person, that might seem anti-social. It’s all about perspective of course, but worth noting. I think that anti-shithead is a good kind of anti-social. We can’t always be fully inclusive, especially of someone who wants everything to go their way. We’d be even more anti-social if we let a singular person have full reign over the things we create and build.

HCI research also enables alienation

The real idea (the spicier one) that I wanted to write about here is how human-computer interaction research is actually, unfortunately, a precursor to something that AI also does: it alienates people.

When you have good and genuine builders and makers, they’ll rightfully want to connect to the real people who are affected by the systems, tools, technologies, and policies they are building. However, as I know full well from accessibility work in particular: many people do not want to actually work with people with disabilities (or give them any power).

And here comes HCI research. HCI research, practically speaking, often operates in a way where power can be maintained by the builders and makers because knowledge-production is made into a separate, “neutral” activity that is conducted by specialists. Does it always work this way? No. But does it mostly? I would argue, yes.

HCI researchers are specialists. They stand between people, such as people with disabilities, and the system owners, technology builders, and so on. We have become mediators of knowledge, but that knowledge is only extracted unidirectionally. And this is, by its asymmetrical structure, a power imbalance too.

This pattern, of a thing put between things in order to create power imbalances, is fundamental to many systems of power: authoritarian governments, bureaucracies, HR in the workplace, and capitalism in general. The best capitalist ideas aren’t ones where someone makes something new or “wins” a rat race beating people in a fair fight, but where a gate can be built between things where people otherwise would have fluidly moved freely. Crossing a bridge? Tollbooth. Healthcare? Insurance. Spending money? Credit cards. And so on. This separation between people and people (and also between people and things) is alienation.

HCI researchers interact with people and translate what they find to people in power because people in power don’t want to do this work, and certainly don’t want to do it themselves. In this way, HCI research alienates builders and makers from other people and alienates people from the builders and makers.

And lo and behold: the once-human-to-human practice of HCI research has now been ursurped by prompting and multi-million dollar startups that claim to have crowds of human-like (and post-human-like) agents who can scale knowledge extraction, albeit via simulation. The core idea claimed is that this method is even better than working with humans.

And because HCI research is simply imagined and envisioned as a vehicle for extracting knowledge from an “other” and bringing that into the work that a person or company is doing for themselves, then a way to get more knowledge, in a bigger format, even faster is obviously an improvement. HCI research, in this framing, was a stepping stone towards synthetic knowledge production at scale.

HCI research was set up as a fragile operation from the start: we didn’t have the power. We acted on behalf of the people with budgets and money and so on. And the people we work with certainly rarely have any power either. And because our job was to act as a knowledge-extractor for systems of power, HCI research in the workforce is at risk of being prompted away.

But… is knowledge extraction really all that HCI research is actually good for? And even if we stick to the core framing of research-is-for-knowledge, we need to interrogate who the knowledge should be for. And in order to do that, we need to pay close attention to the framing and motivation (and even funding) that underpins all of HCI (and especially our own personal work and agendas).

What do we do?

Well, there is a broad and deep collection of sub-disciplines, practices, methods, and people who don’t do HCI research in this purely extractive company-edifying way. Are these people and methods employed in industry settings? Not many that I know of. But do they demonstrate a more-hopeful vision of what HCI research has been and should be about? Yes.

One of my favorite introductions that I suggest folks read is a primer on disability justice in HCI research by Sum et al. It’s fantastic and just the absolute tip of the iceberg on this topic. Disability justice, as just one example of many, completely flips and reframes what should motivate our collective work.

In my humble opinion, the core way that HCI research should operate is disruptive. It should be a way for people who may not have power to now have a say, a seat at the table, and perhaps even budgets or staff who work to make their lives better. Synthetic agents don’t require any power at all! Which is why those ideas, as replacements for HCI research, are deeply perverted, anti-social, and alienating.

HCI research can and should operate as ways to hold people accountable who have power, disperse and distribute power, and even enable more radical alternatives to building and making things. HCI research isn’t really about making better products. Products are always a proxy for some other function or process in our lives. So HCI research should be about living better lives. HCI research should be about the redistribution of knowledge for everyone, and not just a concentration of knowledge handed to corporations and systems that already have immense resources.

Do what I call “robin-hooding” HCI

And yes, in my own self-awareness I recognize that HCI research is essentially an arm of the tech industry, even in academia! My own funding has largely been thanks to companies (Adobe, Apple, Highsoft, and so on). But if companies fund me to do my work and they’re happy with the outcomes, isn’t it alright if I am leveraging those funds to do a little bit of care-filled infrastructure-building and maintence work, too? Or to put this in analogous terms: If we could convince Microsoft to fund public roads, no strings attached, I’d argue we have won some decisive battles.

I call this style of reseach “robin-hooding:” you convince budgets that would otherwise go to something harmful, problematic, profit-serving, or virtually useless to go towards work that is necessary, good, and foundational: care and repair. In my own work, my vision is for every data visualization library to become accessible someday. I left VISA to do this work. I did this during my PhD. I am still doing this work today. I believe that every person who lives with a disability and has to interact with charts and graphs that exclude them, deserves better. So for me, I need to find ways to (as I say) “get away with it:” that is, convince people in power to pay me (and others with disabilites) to fix their broken systems.

And on accessibility and interfaces: The tech world around us is a new kind of infrastructure. It isn’t publicly-funded roadways. It is massively silo’d, closely regulated little fiefdoms of privatized software and services. Unless we have a radical revisioning of how all of tech works today, each individual company has to fix their own broken roads (figuratively speaking). And I do this work by getting money from tech companies and then contributing (as best as I can) through open source channels.

There’s plenty left to do! Plenty more robin-hooding to be done. And for me, I focus on care and repair. Perhaps you, my reader, imagine other futures and outcomes and dreams that you want to pursue. I’d encourage you to dream of how HCI research can help you along that path.

HCI research is a fundamentally social practice that has been co-opted to enable anti-social systems of power. Time to take that back and return to the anti-shithead, knowledge-for-all roots of our work.