The December 2022 issue of IEEE Spectrum is here!

Close bar

Women Less Likely To Be Shown Ads for High-Paying Jobs

Study shows gender discrimination in targeted advertisements, cause unknown

2 min read
A woman between two men.
Photo: Getty Images

Men are more likely to be shown online ads for high-paying executive jobs than women, according to Carnegie Mellon University researchers. The finding is a result of experiments with simulated user profiles to analyze targeted advertisements served by Google’s DoubleClick ad network on third-party websites.

The study shows that gender discrimination in the world of targeted ads is real. But the researchers don’t know who or what is responsible, Anupam Datta, associate professor of electrical and computer engineering said in a press release. Datta and his colleagues presented their results at the Privacy Enhancing Technologies Symposiumin Philadelphia last week.

The researchers have developed a tool called AdFisher. The tool creates hundreds of simulated users and then changes their preferences or online behavior so that the researchers can study the impact of those changes, such as a change in ads that the users receive.

To study the effect of gender, the CMU team created 1,000 virutal users—half female and the other half male—and had them visit 100 top employment sites. The male profiles were much more likely to be shown ads for a career coaching service for executive positions paying over $200,000. The Google ad network showed this ad to the male users more than 1800 times, but only about 300 times to women.

It’s hard to pinpoint the source of the discrimination, the researchers say. Google’s ad targeting system is complex, basing ads on factors such as personal information and browsing history. In the study, though, the user profiles were fresh and behaved the same way, so the only differentiating factor was gender.

It could be that advertisers choose to target the audience they want to reach. Or it could be unintended results of machine learning algorithms that drive the ad recommendation engines. But those algorithms are programmed by humans and are programmed to learn from user behavior. So they could be reflecting real-world prejudices.

Google’s had a string of embarrassing moments with its algorithms. From The Verge:

Earlier this month, for example, Google was forced to apologize after its new Photos app tagged pictuers of black people as gorillas. And in April, researchers found that the results for a Google image search for the term "CEO" were only 11 percent female, a figure that compares poorly to real life stats, with female executives making up 27 percent of US CEOs.

Much more sleuthing of these online targeting systems is clearly required before researchers can draw any big conclusion.

Datta said in a press release that there’s no evidence that Google is doing anything illegal or that it violates its own policies. He is now working with Microsoft Research to get a detailed peek inside Microsoft’s ad ecosystem. 

The Conversation (0)

Why the Internet Needs the InterPlanetary File System

Peer-to-peer file sharing would make the Internet far more efficient

12 min read
An illustration of a series
Carl De Torres

When the COVID-19 pandemic erupted in early 2020, the world made an unprecedented shift to remote work. As a precaution, some Internet providers scaled back service levels temporarily, although that probably wasn’t necessary for countries in Asia, Europe, and North America, which were generally able to cope with the surge in demand caused by people teleworking (and binge-watching Netflix). That’s because most of their networks were overprovisioned, with more capacity than they usually need. But in countries without the same level of investment in network infrastructure, the picture was less rosy: Internet service providers (ISPs) in South Africa and Venezuela, for instance, reported significant strain.

But is overprovisioning the only way to ensure resilience? We don’t think so. To understand the alternative approach we’re championing, though, you first need to recall how the Internet works.

Keep Reading ↓Show less