To think Microsoft Teams is a sexist load of male wankery

[GetUpAgain]

Maybe it's just my experience, I work in STEM, but I fucking detest Teams. And now Zoom has bad press I'm being pressured into more and more Teams wankery.

In a meeting, you only see the faces of the most talkative people. As men tend to dominate meetings this makes women (who are already less-heard) less visible.

The facility to share documents and collaborate in changing them just leads to 50 million wrong versions of the almost same thing. With, ime, men cocking up what is meant to be finished, because they want the last word.

'I will share my screen' = I didn't prepare properly and will now waste your time mansplaining shit you already know.

And then Microsoft sends me this emails telling me how smart I am working. It can get to fuck. That's not how you measure productivity.

Can someone please just hack it, let's have a scandal that means we don't have to use it anymore.

Gotta say I really don’t miss the corporate world. I now work in the transport sector and, despite it being much more male dominated and non-pc, I actually like being able to call a spade a spade.

The worst company by far I’ve worked for was one which made a huge fuss about being ethical. It basically meant that ‘banter’ was strictly forbidden and I was actually disciplined for using the word ‘crap’ in an internal team meeting with just my boss and my administrator. Everybody seemed on tenterhooks about saying the wrong thing, whereas with my current employer we’ll address people who act like knobheads but by and large the working dynamic is better as people aren’t putting on a visage.

I like the fact that my boss can ask me what the problem is and I can literally say “he’s a lazy prick, that’s the problem.”

Well, I never thought Bing would win at anything !

I must admit, I actually expected hands (because of nurture connections in language bias and handcare advertising being nearly all about women) to be predominantly female.

just doubt that Teams, in particular, is to blame for a problem that women in business have been facing since times immemorial

It is however software versions being released in 2020 which still reinforces male stereotyped behaviour which we know is a problem. It doesn't have to to do that, that was a choice by the designers.

The OP's fundamental problem is the team - her colleagues' behaviour, poor meeting management and inherent sexism in both but the way Teams works does magnify that behaviour.

It is depressing how many people honestly believe software and search algorithms are somehow magically bias free when every piece of evidence we have proves the opposite. Its a huge issue for AI in particularly, especially when such systems are increasingly used for medical diagnosis, prediction/assumptions of criminal behaviour and fraud and workplace/financial decisions.

I am afraid that any software that would put women in the spotlight even when they are not saying anything would actually be more dentrimental to their development and confidence

Why would that be the only alternative?

The images in video calls could be current speaker plus a randomised selection of the others and change them every few minutes. That is the point (I think!) which the OP is trying the make, that choices were made in design and development which keep the loudest people as the primary focus e,ven when they are not speaking, rather than share the focus between attendees. Nine windows may help for smaller meetings but the underlying behaviour of prioritising the most frequent interjectors is still there as far as I can see.

I agree with PPs recommending "Inivisible Women" - in general this is a much bigger issue than people are acknowledging.

I don’t want to see randomised pictures - I want to see those speaking

(otherwise might as well use Skype )

But is it just the contents of the internet or is the search engine biased?

Search engines contain huge numbers of decisions - those decisions reflect the biases of the team developing them from specific rule based decisions to selection of of the subsets of data supplied as "learning" material.

So as a simplistic example the team at Amazon who decided to teach the recruitment AI solution based on their CVs/board CVs selected candidates which were overwhelmingly male, white and advantaged. Its particularly noticeable on language translation searches where the defaults are always male and searches are even "corrected" to male defaults. In an era when "use Google" is the stock response to every question this does matter.

It is genuinely difficult to avoid building systems which enforce prevailing biases, its completely impossible when people don't even believe the problem exists.

I don’t want to see randomised pictures - I want to see those speaking

What I actually said was:

The images in video calls could be current speaker plus a randomised selection of the others and change them every few minutes.

Teams is incredibly annoying in that you can only see who's talking - and you're right, that's usually one or two people - and they are invariably men.

I despise teams as the next new thing that I see has no purpose in my life and worker but I keep getting it shoved down my throat and told to join- I do t see it as a bastion of male strength and female disempowerment, but an annoying phase which I'm hoping will bugger off.

I came back on to make a similar points to @C8H10N4O2. I'm not saying that Teams is gender biased - we cannot possible claim that based solely on our individual and anecdotal experiences. But it absolutely could be, and it is vitally important that we understand that it could be.

As the PP says, technology like this typically learns by the data sets it is trained with and then by ongoing reinforcement (this is a generalisation, but broadly true). If, through its training, it is marginally better at recognising male voices, it may be that when two people start talking at once, it switches to the male speaker more frequently than the female speaker. This bias may be tiny - undetectable by us as individual users - but scale it up and you may well have the situation that globally we are using software that reinforces men speaking over women.

Note that this is just an example of what might be happening - it may be, it might not, it might be that there are other similar things happening instead. But we should be challenging it, constantly.

The fact that people don't understand the seriousness of the issues at play here makes me thing that Stephen Hawking was right and that we are sleepwalking into a truly awful future.

In a meeting, you only see the faces of the most talkative people. As men tend to dominate meetings this makes women (who are already less-heard) less visible.

You can override this by pinning people you want to see.

Teams isn't sexist. People are sexist.

Bold fail. If only there was an edit button.

And it is so important I will say it again - if you don't understand the issue then please, please read Invisible Women by Caroline Criado Perez‎. And make everyone you know read it too (especially if they are software engineers).

It is depressing how many people honestly believe software and search algorithms are somehow magically bias free when every piece of evidence we have proves the opposite. Its a huge issue for AI in particularly, especially when such systems are increasingly used for medical diagnosis, prediction/assumptions of criminal behaviour and fraud and workplace/financial decisions

Exactly.

@fuckinghellthisshit
So much so that he now mutes several men on his team for the majority of the meeting

Do you really think that this is the best way to deal with it?

You can override this by pinning people you want to see

Most large corporate deployments of Teams (and similar) restrict the range of options. This also doesn't change the default behaviour, which is the fundamental issue here.

As a side note for fellow Criado- Perez fans, have you met a man who has read it?

@C8H1... But that is how Team works. Shows the current speaker, the previous speaker and two other random people from the group. At least that is how we set up ours.

And putting forward people who don't speak by pinning them can feel like pressure to them to contribute when they are less prepared/may not be comfortable doing so, thus making them less confident. Obviously this does not apply to those men and women who are perfectly comfortable in the setting and have no issues claiming space/time to make their opinion known.

If, through its training, it is marginally better at recognising male voices, it may be that when two people start talking at once, it switches to the male speaker more frequently than the female speaker.

Is that proven or are you just proposing it as a potential factor? If anything, I think female voices are generally easier to pick out as a higher pitch is more distinct among everyday rumbling, shuffling and bassy sounds. I'd always choose a female satnav voice as it's clearer above the road and traffic noise. That's one (of several) reasons why babies find women's voices preferable and indeed perceptive men will often raise their voice pitch when communicating with very little ones.

@EatsShootsAndRuns It was an observation, that's all. DB is a director at a major IT company - personally I think the way they handle most things is utterly shit.

@Piggywaspushed DS is 16 and read it, DB some bits of it encouraged by DS.

I hope it’s not too rude to slightly sidestep some of the OP’s initial post, but I switched over to BlueJeans after the Zoom furore (also in STEM, lots of friends in software) and it works well. Absolutely no chance we’d have started off with Microsoft Teams though, so can’t directly compare to that.

Glad to hear it! I have no idea how to shorten your username without sounding terribly insulting fuckinghell !

At least that is how we set up ours

That is the key point - what is the default behaviour, why should it need tweaking to reduce bias?

Behaviour will vary a bit between versions (and types of software - Teams isn't unique after all), it will also vary depending on mandated configs from complanies. However most work fundamentally on a "Lease Recently Used" model which inherently favours the loudest and most frequent interjectors (even if theire contributions are rubbish) being on screen where there are less windows than people.

Is that proven or are you just proposing it as a potential factor?

Honestly, read Invisible Women if you haven't already done so. This phenomenon is long documented and researched but that book is a very accessible compilation of good examples with all the background references.

Its not just sex bias which is an issue either. Ethnicity, social class, accents, age and conformance to local concepts of attractiveness can all be factors of bias built into systems.

This is an infamous example of colour bias built into technology and relatively simple software, actually operating in the offices of a Silicon Valley hi tech:

So did the team think "We will make a soap dispenser which only works for white people"? Highly unlikely (I hope) but they did build a solution for their unconscious view of a default person and simply didn't consider another type.

Another current but non technical example of defaults would be PPE provisioning for women who despite dominating the health workforce have to use PPE designed for male bodies.