Featured Stories

Are You There Google? It’s Me, a Woman.

Diverse voices in tech won’t solve anything if they literally can’t be heard

Danielle Morgan
Mar 12 · 7 min read
Illustration: Danielle Morgan

Listen to this story

--:--

--:--

“You’ll just really have to know your stuff.”

I shifted in my wobbly red chair. If you’ve ever given a talk at a well-funded but underpromoted university event, you know the kind of chair I mean. The grooves of the plastic didn’t fit around my body, and the seat was digging into my lower thighs. The man I was chatting with was leaning back comfortably in his chair with his arms crossed.

“Why?” I asked him, and he turned his rimless glasses toward me (if you’ve ever given a talk at any university event, you know the glasses I mean) and warned me about the audience.

Every time I give a talk, the other speakers always like to warn me about the audience. Usually, they let me know that there are a lot of men, that these men are from the engineering department, and that I’ll just need to really know my stuff.

I think the other speakers are genuinely trying to be helpful and supportive. I am frequently mistaken for a teen. If someone asks me if I’m finished with school, I know they aren’t talking about university. I’ve never purchased alcohol without someone laughing at the attempt. I was pulled over at 26 because a police officer thought I was too young to possibly possess a valid driver’s license. So, the comments aren’t coming out of left field, but they’re still pretty unwarranted.

The idea that women need to know more, do more, and be more in order to be taken as seriously as any average (or even below average) man is obvious and well-documented. We all know it’s true, but people rarely say it out loud. I kind of appreciated this man’s straightforward acknowledgment of it. Kind of.

I wanted to say, “I don’t need to know any more than you do. If these people aren’t going to take me seriously because of the way I look, they won’t take me seriously even if I do show them that I ‘know my stuff,’” but I didn’t say it. Instead, I narrowed my eyes slightly and nodded.

The chair situation bothered me more than the comment, honestly. I sat through this man’s entire talk, while perched on the edge of that horribly uncomfortable chair. It was like being forced into ill-fitting clothing for a half an hour before going onstage. (Or it was like being forced to sit in a chair built for a very tall, sinewy man for half an hour before going onstage.)

When it was my turn to speak, the microphone had trouble picking up my voice and I fiddled with it for a minute. Let me be clear: The microphone at this event couldn’t pick up my voice while I was attempting to talk about technological bias against women’s voices. I nearly burst into laughter.

I watched as 75 percent of the audience glanced at their phones. No one was paying any attention, no one could hear me, and no one cared if I “knew my stuff” or not.


Over the last few years, I’ve noticed a lot of hype about technological bias in the news. I’m seeing more headlines like “Research Shows Gender Bias in Google’s Voice Recognition” and “When Technology Discriminates: How Algorithmic Bias Can Make an Impact.” I’m glad people are talking about it, but these articles treat the issue like a problem that can be solved through inclusion alone.

We can’t just keep forging ahead without understanding how we got here.

Almost every article I read ends with the assertion that technological bias can (and should) be fixed by bringing more women, more accents, and more voices into the tech world: Include more! Let them make more decisions! Pay them more!

This is great. We need to include more people. Obviously, we should allow more people to make decisions. For the love of god, pay people fairly. This will solve a lot of problems, but it’s not going to solve technological bias, and it won’t make people more comfortable in the environments we’re suddenly “welcoming” them into.

We can’t just keep forging ahead without understanding how we got here. We can’t just keep hiring more people to work in the same biased environments and with the same biased systems. To even attempt to fix these issues, we actually have to understand where these problems came from.


Here’s where a small example of technological bias originated: In 1898, a man named Valdemar Poulsen used piano wire, an electromagnetic trolley, and rewired telephone parts to magnetize a length of wire with his voice. This was the first documented experiment with wire recording, and it was the first time someone heard their own voice played back to them from recorded wire.

I wanted so badly to replicate Poulsen’s experiment and hear my voice on a similar wire recording. Unfortunately, even after dedicating an entire year of my life to re-creating his experiment, I’ve still never recorded my natural voice on wire. If a woman had been in the room with Poulsen while he was experimenting, she probably wouldn’t have been able to experience it either because, according to my research (10 points to the kindred spirits who get that reference), the contraption was unable to record higher pitched voices.

My (female) research partner and I rewired old telephone parts, strung piano wire, recreated the entire experiment as precisely as we could, and we screamed into those telephone parts until our department complained about the noise levels. We wrongly believed that if we screamed louder, surely the mechanism would pick up and record our voices. It wasn’t until we decided to experiment with the pitch of our voices that we got somewhere.

We spoke in deep, husky Sean Connery voices and the thing worked immediately. Previous weeks of tweaking the mechanism were for nothing. Our replicated wire recorder worked just fine. It just couldn’t pick up our shrill lady screams.

To test this, we asked our male colleague to try out the recorder. After a moment of him speaking normally into the recorder, it worked. My partner and I stood there watching him nonchalantly record his voice.

It wouldn’t have mattered how many women you put in the room with Poulsen or how many you put in the room with us. The physical technologies were biased against higher pitched voices. The presence of more women using the same pieces of biased technology was not going to change that fact.

Yes, some women have lower pitched voices and maybe a woman with a lower pitched voice could have helped us figure it out. Yes, some men have higher pitched voices and wouldn’t be able to use the mechanism. Yes, this experiment is extremely limited (and flawed) and does not address how the mechanism would work with different accents and body types. It does, however, show that wire recording was probably built primarily (or solely) for lower pitched voices.

Every single piece of recording technology and voice recognition software we have today originated from these biased experiments. It’s really not a mystery why people come up against technological bias, and we should stop treating it like one. We should also stop suggesting that simply allowing people with different kinds of voices into offices and labs will fix it. Technological bias is as old as technology itself. Putting a bunch of diverse people together to work in biased environments with existing technologies and algorithms that are biased against them will solve nothing.

It doesn’t help to simply allow people to speak at your events, to test biased technologies with accents and higher pitched voices, to let people work beside you in an environment that was built specifically to exclude them. Instead of simply welcoming people, we must work on rebuilding biased algorithms that privilege certain voices, reworking physical structures and spaces that privilege certain body types, and dismantling any system built for one kind of person.

My complaints about chairs and microphones at cushy university events are ridiculously small compared to the discomfort and loneliness some people must feel when confronted with algorithms and structures that were built to keep them at a distance. I’m not saying that dismantling technological bias is a simple task, and I’m not going to pretend that I know how to fix it. But I don’t think the answer is to include as many people as we can as quickly as we can.

I want the world to be deeply diverse, but I also think it’s important and necessary for people to feel comfortable and heard in that world. Allowing people to speak at your event isn’t the same as creating an environment that allows them to be heard. Allowing people with different voices to use and work with voice recognition and recording technology isn’t the same as rebuilding these technologies so they work better for those voices. I think we can do more than just allow people to exist in a world that wasn’t built for them.

I finished my talk and everyone applauded. I walked off and no one talked to me about my work for the rest of the day. The other speakers and I ate sandwiches together, I engaged in polite conversation with a very comfortably seated redheaded man about his research, and then I headed home.

Now, every time I sit on my bed, ask the Google Home to play Gilmore Girls on the TV, and she says, “sorry, I didn’t catch that,” I giggle at my own reflection in the dark TV screen.


Research on Valdemar Poulsen and his wire-recording experiments was done in collaboration with the University of Victoria’s Maker Lab.

Danielle Morgan

Written by

I have feelings about that. daniellemorgan.org Instagram: daniellemorganstudio

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade