Social media algorithms 'slip' easy money adverts to poor young people

This has been shown by a pioneering study from the Universitat Pompeu Fabra (UPF) in Barcelona, which for the first time analysed how socio-economic status and gender affect personalised advertising on TikTok and Instagram.

Among other things, the study showed:
15% of lower-class young people receive adverts for risky financial products (fast loans, crypto investments, etc.), compared to only 8% among their upper-class peers.

The work is published in the journal Communication & Society.

To whom and what algorithms show

The study is based on an online survey of 1,200 young people from Catalonia between the ages of 14 and 30. Participants were asked to rate promotional videos that spontaneously appeared in their TikTok and Instagram feeds.

It turned out that young people from less affluent families were the most likely to see adverts:

• risky financial services (quick loans, crypto-investments, etc.),

• betting and online gambling,

• online gambling,

• content promising "easy money" and "flexible work without requirements".

For young people from more affluent families, the most frequent ad unit is travel and leisure.

The main author of the paper, Carolina Sáez from the CAS (Communication, Advertising and Society) research group, conducted the research under the supervision of Mònika Jiménez. Isabel Rodríguez de Dios (University of Salamanca) also participated.

"If we look at adverts promising to climb the social ladder, the differences become huge," notes Saez.

Promises of "easy success"

Young people with fewer resources are much more likely to think they're getting adverts:

• easy money in digital businesses - 44 per cent vs. 4 per cent for the upper class;

• "affordable" jobs with no requirements - 39% vs. 4%;

• investments with "quick returns" (e.g. through cryptocurrencies) - 33% vs. 4%;

• instant part-time jobs and earning money from your phone - 27% vs. 3.5%;

• quick loans with "no strings attached" - 21% vs. 3%.

The researchers conclude that it is the more vulnerable young people who are being targeted by adverts that play on their desire for a social leap.

Gender stereotypes in digital advertising

Differences by class are particularly pronounced among young men. For example:

• lower class boys/men are twice as likely to see online gambling adverts than their upper class peers(22% vs. 11%);

• for girls, the spread is much smaller: 6.7 per cent vs. 5.6 per cent respectively.

That said, gender differences in advertising in general are very pronounced:

• girls are much more likely to see fashion(50% vs. 13%) and cosmetics/beauty(71% vs. 28%) adverts;

• girls see more than three times as many parenting-related adverts(16 per cent vs. 5 percent);

• boys are more likely to see adverts for sports(54% vs. 26%), online games(46% vs. 23%), technology and electronics(32% vs. 15%), and energy drinks(10% vs. 4%);

• and three or more times more likely to see auto(16% vs. 6%) and alcohol(10% vs. 4%) ads.

How algorithms "guess" class and gender

Formally, European data protection legislation restricts platforms from accessing sensitive information. However, the gigantic amount of behavioural data allows TikTok and Instagram algorithms to indirectly reconstruct:

• economic status,

• education level,

• origin,

• the user's professional status.

Saez tested this hypothesis:

• participants anonymised their postal address during the study,

• this data was compared with the Índex Socioeconòmic Territorial (IST), the official territorial socio-economic index of the Catalan government, which takes into account employment, income, education, etc. The final " class" according to the IST largely coincided with that of the user;

• the resulting "grade" on IST largely coincided with how algorithms target ads.

"Algorithmic advertising commercially exploits this information and plays on the desire of young people from disadvantaged backgrounds to climb the social ladder," Saez warns.

Children see alcohol and gambling adverts

Another worrying finding: teenagers aged 14-17 reported seeing adverts for:

• alcohol,

• gambling,

• electronic cigarettes,

• energy drinks -

even though European regulations should protect minors from such content.

The authors conclude:

• regulation and control of the use of AI in online advertising aimed at young people needs to be strengthened;

• in parallel, digital literacy of teenagers and young adults should be developed to increase their critical attitude towards advertising and algorithms.