How I changed my gender

On social media

Anita Verő
9 min readMar 31, 2021

The story

In the past few months I got annoyed with Facebook because it flooded my feed with sponsored ads almost solely about yoga pants, other fashion products, clothes and jewelry, and nothing else, even though I have no interest in buying any of these. (Yes, I probably spend too much time on Fb since the corona situation got dire in Hungary.) After a while I decided to try and train the recommendation algorithm and started blocking these ads but nothing changed. One day I clicked on the “Why am I seeing this ad” link to see what they come up with, even though I know that

  1. They probably use a type of algorithm which cannot necessarily explain it’s decisions in human language (see next chapter).
  2. Even if they use a more explainable version, I do not have any guarantees for getting the true information.
Sponsored ad while being “female”

Nevertheless, I decided to go down the rabbit hole once and for all. (Yes I became a bit preoccupied with this, but I can rationalise it by calling it “doing research”.)

On a yoga pants ad the explanation was that the company who was advertising is targeting people who are interested in fashion. In the same window there was a link where I can set my “interests” that had been somehow generated by Facebook beforehand. I went through them and deleted everything related to fashion and only left things related to science, AI and politics…

Nothing really changed. To be fair, I did get fewer yoga pants, but I got my fair share of bikinis, clothes and jewelry in return. This was a point of no return. I clicked “Why am I seeing this ad” on a bikini ad when finally I felt there was a moment of truth in Facebook’s communication towards me:

“This company is targeting females between 18 and 35”.

I knew right away what I just had to do. I went to my profile, where indeed my gender was set to “female”. There is a drop down menu with more than 70 available gender labels. I decided to select all of them (didn’t wanna get FOMO). Unfortunately, there is a maximum limit of 10, so I left all the more neutral ones leaving me with: Agender, Neither, Other, Gender Nonconforming, Not specified, Genderqueer, Gender Identity, Gender Questioning, Non-binary and Gender Fluid.

This worked like magic, but not the way I would have thought. The bikini ads had been replaced by expensive cars, petrol stations, business tips on how I should handle my employees (apparently I immediately got promoted) and a weird ad on microdosing LSD. I basically became a stereotypical man by default, in the eye of Facebook.

Sponsored ads while being ⚧⚲☿X etc.

Since then, sometimes I get ads which are even relevant, although I do not click on any of them and will probably search for (or develop) a better ad blocker which blocks sponsored ad posts too, not just banners. But there were things on science talks and books which could even potentially interest me.

Bias in AI

In the past few years researchers started actively questioning the social consequences and ethical issues of Artificial Intelligence.

There are numerous aspects we could talk about from data bias, fake news and echo chambers, through company incentives and potential regulations, as well as longer term effects on the job market and the way we look at human creativity.

But now, I would like to focus on one particular aspect that any of us as users of this technology should be aware of and things we can do to protect ourselves from its harmful effects today.

This story above is only a funny anecdote, but it is an illustration of the problem of data bias and models which can reinforce these biases by parroting back all the stereotypes they learn from data. This happens for several reasons from the unrepresentative demographic of content creators on the internet to general social biases. These effects are carefully summarised with loads of citations in the latest work of Timnit Gebru, Margaret Mitchell (and others from Google who didn’t dare to reveal their names). They got fired from the lead of Google AI Ethics team, as a result of wanting to publish this paper, which may reveal a bit too much about the responsibility of huge data driven companies, such as Google.

Bias feedback loop

The algorithms aggregate the information they are fed with, and on social media platforms they throw it back at us, creating a feedback loop where they influence us as much as we influence it. We are in a constant conversation with a system we don’t really know how to communicate with and are often not even aware that we are in communication.

And even if we are, the interpretation of the algorithm’s models of an individual is difficult if not impossible for any human, including the developers. The very reason we use these algorithms is that they process large amounts of data much faster than we can and make decisions based on their own “ideas” about it. We can try and query how these models see the world but it will always be a simplified, and potentially misleading explanation, similarly if someone asked you why exactly you picked strawberry ice cream instead of chocolate at a given moment. (There must be some internal processes that led to this result but many of them aren’t available for us to verbalise.)

Okay, but what do you have to do with this?

…you may ask. Tristan Harris mentioned the need for a “self-awareness enlightenment” in his influential TED talk.

The first step would be to notice that we are all impressionable, even if we are aware of these effects.

The second step is to start noticing when and how we are influenced, what our weaknesses are and how other humans or an algorithm can play onto that (it is usually related to some deeper fears).

This is hard emotional work which doesn’t happen overnight, but in the meantime we can take some easier steps to take control over our mental diet on social and other media.

There are some useful tips here, but I would also highlight some and add my owns:

  • Switch off autoplay on video players such as YouTube.
  • Use ad blocker.
  • Switch off notifications for anything but private messages or what you find really important and beneficial for you.
  • Take days off social media.
  • General mental health “workouts”.
  • Follow people you disagree with.
  • And maybe try and change your gender, age or other settings on Facebook, to see what worlds other people might live in. ;)

More on companies’ race for our attention and their use of efficient psychological techniques to keep us on their platform, by carefully timed notifications, highly emotional contents etc. are powerfully (if a bit cheesily) illustrated in the movie called Social Dilemma.

Bias in real life

Despite the clickbaity title I did not really change gender. Or did I?

From: https://marketingland.com/are-brands-behind-the-times-when-it-comes-to-gender-stereotypes-in-ads-255997

While growing up I wasn’t really aware of my gender. I was fortunate enough that I wasn’t really forced into many boxes, I had dolls as well as toy cars and lego. I performed well at school, I was interested in sciences as well as in literature later on.

The first time I started noticing weird things in people’s reaction towards my way of existence was around high school. I wore girly outfits, nothing extreme, usually some kind of skinny jeans, low high heels and makeup. Throughout these years I found myself in the same situation over and over again. I was pretty socially anxious, so it was hard to get to know me, but it did happen sometimes, and when it did, people usually told me they were surprised I am “so smart” and interested in philosophy, science and other things, because they thought I am dumm, based on the way I look. After a philosophical monologue of mine one person even told me that with these kinda thoughts I should wear boots and a hippie bag. My first reaction was that these people are out of their minds, I was not able to comprehend how my appearance is related to my intelligence.

Later on I was often told by guys that I am like a “guy” because I have a dry sense of humour and because of the “way I speak”.

My confusion culminated at university, where gender disparity hit me first. I started with studying German linguistics and literature with a ~95% women student body, then after a year I moved to studying Computer Science with the exact opposite gender ratio. (This was around 2009 and I knew nothing about non binary notions or even the word gender.)

For a while, as a desperate attempt to stop being put into boxes based on my appearance, I started wearing baggy jeans and flannel shirts (like a true IT guy) and stopped wearing makeup for a while (for which I got criticised by acquaintances who knew me before). I was “finally” wearing boots as well.

I even went through a phase when I internalised being “different from other girls’’ and thought I “cannot connect with girls only guys”, which I am only shaking my head about now. Fortunately, I soon dropped this entirely, which gained me many intelligent, interesting and kind women friends over the years, without whom my life would be much bleaker.

This confusion about outfits may seem rather funny and pretty harmless, but things got more serious when certain academics during my Master’s (when I already worked in AI), told the whole group, that women are better at administration, worse at navigation and spatial imagination (btw I got 3D modelling exercises right always among the first people at computer graphics lab sessions). They expressed such rhetoric on a regular basis during group meetings (where I was the only girl) and not a single time did anyone (including me) say a word.

When I was applying to Cambridge and Edinburgh for a PhD one professor told me I would surely get in, since I am a woman (and they must favour women in Western universities these days — this was in Budapest, Hungary).

I did get into both places with full scholarship offers and I chose Cambridge in the end.

I would lie if I told you all this didn’t get into my head. Of course I had impostor syndrome since I entered Computer Science. I don’t think this is mainly because of my gender but these feedbacks from outside definitely didn’t help.

I want to emphasize that I do not blame any individual for this phenomenon. As I mentioned earlier, I myself discovered pretty wild prejudices in my own thinking and I know how easy it is to fall for these ideas. No one is immune to this, regardless of background and it can be directed outwards and inwards too and even at the same time. I also know that even if it’s hard, people are able to change their thinking with time.

I told all this just to demonstrate, facing people vigorously wanting to fit me into their tiny boxes, is very much not new to me. Only this time it is done by an algorithm aggregating all these opinions and throwing it back at me automatically.

I will definitely try to be more conscious about my media usage in this regard and in general, and I would encourage you to do the same.

Let’s play a game and try to take more control over what ideas are fed into our own minds and maybe try and exploit this technology to peek into others’. How about we all change our gender labels around on Facebook and shake up digital and our own human models of ourselves as a start. ;)

--

--

Anita Verő

I do Machine Learning, code, sing and sometimes write blog posts.