What do you do when your video conferencing software is sexist?
Sometimes, I interrupt people in meetings.
I try not to, and I’m working on being better at noticing and calling myself out, and at noticing when others do it and calling them out too.
But I’m definitely not perfect at it.
This is not ideal: I’m a lucky and privileged white dude running a company. I want everyone in meetings to contribute, to feel heard, to be heard, to be safe.
So I keep working on myself. And I try not to ask others to change to fit my arbitrary way of expressing myself. I don’t want you to have to call me out, to speak up louder, to interrupt me back. But recently I’ve hit an interesting edge case and I wonder if anyone else has come up with a good way of dealing with it.
What’s this got to do with video conferencing?
I imagine most of you have made video calls. A good number have most likely been in group meetings conducted over video. We have a weekly leadership team meeting spanning three offices and eight timezones and attended by 6 of us: me, three other men and two women.
Both the women are based in the same office as me, and so I’m often in-person with them and joined by our two American office VPs remotely. I’d been noticing that the remote attendees seemed to ask the women to repeat themselves more often than they asked the men. Then: last week both the women were dialling in from out of the office and I realised: it was so much harder to hear them than it was to hear the men who were dialling in. This is not generally the case in-person — those of us sat together in the room can all hear each other just fine.
So I got to thinking: is our hangout software sexist?
I’m sure there is no grand conspiracy here: women may in general talk more softly than men, and they on average have higher-pitched voices. I suspect that the former in particular is straight-forwardly to blame: the algorithms that cut out background noise, and enhance the voice of whoever is speaking no doubt work better when you speak louder. It seemed that this combines to mean that the women cut out more often or had their speech garbled.
What if it’s not just a volume thing, and the compression and enhancement algorithms also work better when you speak in a lower voice?
Is it possible that most of the testing has been done with groups of men / mainly men? [See: how colour film was biased towards white people]. Is it possible that when women have been in the test meetings, they have forced themselves to speak up more than they naturally would, to be heard?
It’s surely possible to imagine an algorithm designed to counteract this? Has anyone tried?
Because the last thing I want to do is ask the women in the meetings to just speak up, to just drop their voice a bit. That’s BS.
But I’m struggling. While I’m not in control of the video conferencing software, what else can I do?
Has anyone else noticed this phenomenon? Solved it? Got any bright ideas?
I’m all ears.
