How Biased Content Algorithms Reinforce Gendered Stereotypes in Young Children

From gendered wall paint to toys and clothing, we enforce gender roles on children from a very young age. However, what we expose our children to during their formative years plays a huge role in shaping their future – especially with regard to their career choices and personalities.

Games, textbooks and social circles – everything influences a child’s vocational appetite. Gendered teaching and games aside, another key factor impacting our younger generation is their reliance on technology. Though the child-screen relationship seeds were sown long ago, the bond between a child and their favourite device has only grown deeper during the pandemic.

‘Games for boys’ and ‘games for girls’

While machines can greatly aid a child’s educational development, there are some areas of concern. Whether you own an Apple or an Android device, just hop on the AppStore or Playstore and search for “games for girls”. The top ten results contain some variation of cooking/baking, make-up, fashion and household chores simulators. On the other hand, a search for “games for boys” will result in a mix of races, puzzles, math and science games.

“The former bias of gendered toys (Barbies for girls and trucks/puzzles for boys) has transcended into the digital space with gendered apps,” says Riju Chandrayana Gupta, senior account manager with Schlumberger, Petroleum Engineer, who has two daughters under the age of 13.

What a child sees in the media (films, television) and apps affects their perception. They create distinct lines in their minds about what boys can do and what girls can do. This eventually reflects in the gender ratio at workplaces, especially in fields related to science. According to the United Nations, women constitute merely 14% of the total 280,000 scientists, engineers and technologists in research development institutions in India.

Knowingly, or unknowingly, we are enforcing expected gender roles in children with the help of smartphones.

“Classification and categorisation of apps are generally not gender-specific,” says London-based data intelligence expert and consultant Kumar Vikram Singh.

However, the desired audience can be targeted by gendered advertising, the colouring of the apps, and so on.

Gendered AI

The way app store search engines work is by a combination of artificial intelligence (AI) and machine learning of the stores based on the Search Engine Optimisation and data provided by the app developers.

Keywords are put in by the developer so that an app falls under a certain category. “The information whether an app is ideal for a girl or a boy is from the SEO side,” Singh adds. Developers use tags to target users who will be the most ideal audience for it. “The main control over the result page and AI is with the Play Store and the AppStore, and not the developer.”

The AI collects data from the public domain – the internet. Gradually, the machine starts inheriting our own biases and stereotypes. If we use more and more cooking content for girls and science games for boys, it will encode that as a fact.

“The results come as an output of a Machine Learning model which is trained by Google,” says Yuganka Sharan, Android lead at Kruzr. These outputs are a result of these inputs – what the developer writes, variables that Google feels are relevant, and previous downloads.

We are teaching these machines and thus, any biases presented are inherently our own.

If a majority of users correlate a cooking app with a girl child via keywords, comments, and so on, the machine then processes this information that cooking games are made for girls. So, the next time another person starts a search for ‘games for girls,’ the AI will flood the result with cooking games.

Kumar says it’s unfortunate that these results are so gender-biased.

But even if the results are skewed, does it really matter in the larger scheme of things? According to some parents and gender experts, it does.

Also read: Trucks For Boys, Dolls For Girls: How Toy Stores Reinforce Gender Roles

“By the time a girl is 10-12, she will absorb from these apps and the media that she is more suited towards such endeavours. We see that a lot of young girls lose confidence in science because as a society, and now these apps, we subconsciously let her believe these pursuits are not made for her,” says Gupta.

What we see is not accidental. Algorithms and machine learning curates the content we see, so can the constantly biased content seep into the child’s psychology?

Yes and no, says Tanushree Deb, a Physics teacher from Delhi. She was one of only four girls in her graduation course, equipment and maintenance in electronics, more than two decades ago. “I still face a bias when I go for an interview. I get told by male interviewers that physics is not for girls,” she says. As a result of this conditioning, many women in STEM fields feel double the pressure to prove that they belong in their well-deserved jobs.

But Deb adds that apps can actually contribute towards breaking gender-roles. A young boy, 20 years ago, could not have openly said he wanted to be a cook or a designer.

Breaking stereotypes

“I never searched specifically for girls when we began, but it showed up in the search suggestions anyway and yes, I noticed the bias immediately,” says Saimah Raza, PhD scholar at Translational Health Science and Technology Institute (THSTI) about her experience while searching smartphone apps for her five-year-old daughter.

She noted how fun games mentioned for boys would be adventure-based, skill-based, technique-based – the likes of construction, racing, puzzles. While for girls it will be on the lines of dressing up, decorating rooms, and baking – “all princess-y and unicorn-y.”

“Nothing wrong at all with either, or wanting to indulge in these, but as a parent who wants to keep things gender-neutral, it’s annoying and quite frustrating,” she adds.

Both Gupta and Raza, themselves women in STEM and mothers, try to monitor what their children watch, read, or play. It’s not a question of control, says Gupta, but very young kids, especially below seven-eight, cannot be expected to make decisions that are always good for them. They try that their children are exposed to a mix of content; whether through games and apps or through media.

Stereotypes at home

A lot of what children learn about gender roles and stereotypes begin at home. Before their access to apps or media, they learn from their family’s behaviour, says Deb. We cannot always blame the app or media if we ourselves are reinforcing gender roles in our children.

She adds what affects children more than media is peer pressure. So, if a gendered game is all the rage among children of a certain age, other kids might be more inclined toward playing it.

“Children are inherently curious and inquisitive, irrespective of their genders. We just need to keep encouraging them, keep presenting them with age-appropriate challenges and puzzles to get them into the habit of solving things,” says Raza.

This divide can only be resolved if we change our own attitudes towards gender bias. As the algorithm experts stated, the AI and machines only learn from our behaviour and in turn, teach our biases to children in the form of gendered apps and media content.

Anwiti Singh is a journalist with a love for science and pop culture.

Featured image credit: Pairplab Chakraborty