Researchers say Facebook's algorithms disproportionately show certain types of job ads to men and women; no evidence of gender skewing was found on LinkedIn
The contrast with LinkedIn matters because both are relevant channels for employment advertising, yet the researchers found different outcomes. It also sets a clear benchmark for the later push by job-search services to remove bias from AI-powered matching systems.
First-order effects
Employers using Facebook for recruiting face evidence that the platform's delivery system can unevenly distribute job opportunities by gender even when the advertiser is not described as selecting recipients that way.
Facebook faces renewed pressure to examine how its ad-delivery algorithms allocate employment ads, while LinkedIn gains a comparative result showing no detected gender skew in this research.
Second-order effects
Recruiters and job-ad buyers may place greater weight on the delivery controls and bias-testing evidence offered by Facebook versus LinkedIn, rather than treating targeting settings as sufficient safeguards.
The findings strengthen the case for job platforms' bias-removal work in matching and ad systems, because unequal delivery can persist beyond an employer's stated targeting parameters.
Third-order effects
Employment-ad compliance is shifting from scrutiny of advertiser intent toward accountability for platform optimization systems that determine who actually sees an opportunity.
If comparative audits become routine, job platforms will compete not only for recruiting reach but also on demonstrable fairness in algorithmic delivery.
The trend: Hiring platforms are being judged increasingly on whether their automated distribution systems produce equitable access to jobs, not merely on the neutrality of advertiser targeting.
The latest audit of FB's ad service shows it's still excluding women from seeing job ads without regard to their qualifications. This is sex-based discrimination under US equal employment opportunity law. Important work from @basilealyos, @korolova & co. https://www.technologyrev…
Facebook's ad delivery algorithms have been known to introduce bias and create echo chambers for job, housing and politics ads for several years. Our latest research shows that for job ad delivery, Facebook's algorithms are not merely biased, but discriminatory under US law. 🧵
2/ We develop a new auditing methodology to distinguish between skew introduced by job ad delivery algorithms that may be explainable by differences in qualifications (permissible by law) from skew due to factors such as the ad platform's optimization for its business objectives.
Facebook is withholding certain job ads from women because of their gender, according to the latest audit of its ad service. https://www.technologyreview.com/ ...
7/ Our work shows that despite commitments in response to prior studies, settlements, & a civil rights audit, Facebook hasn't made visible progress in addressing discrimination in its ad delivery algorithms; and calls for meaningful algorithmic transparency to be mandated by law.
6/ Although there may be a debate about allocating responsibility for discrimination between advertiser and platform when the advertiser asks to optimize for clicks, the responsibility for any discrimination observed in “reach” ads rests on the ad platform.
Eisenstat: I know, you hired me to help you fix how people are manipulating the platform to affect elections. If that's not what you're going to have me do, then I'm out. WATCH #HoldTheLine: https://www.youtube.com/...
Eisenstat: And I understand that the Facebook mentality is, people are just dying to work here, so once they get in the door, they'll just be part of our company, and they'll be thrilled with anything. WATCH #HoldTheLine: https://www.youtube.com/...
3/ The idea is to simultaneously run ads for jobs with identical qualifications but skewed gender distributions in reality. The pairing controls for factors such as competition from other advertisers, allowing to isolate the role of the delivery algorithm in reproducing biases.
Eisenstat: Long story short, they hired me for everything that I bring to the table, and then they pushed me into a corner and silenced me because of everything I bring to the table. They weren't serious about me. WATCH #HoldTheLine: https://www.youtube.com/...
Researchers found Facebook systems were more likely to present certain job ads to users if their gender identity reflected the concentration of that gender in a particular position or industry https://www.wsj.com/...
What is going on? It's 2021. Almost 5 years has passed. Yet new study indicates that Facebook has not resolved its ad discrimination problems since this issue was brought to light in 2016. https://www.technologyreview.com/ ...
5/ The gender skew in delivery persists even when an advertiser chooses to optimize for “reach”, i.e. aims to show their job ad to a broad audience, rather than only to the people likely to click it.
4/ We apply our methodology to Facebook and LinkedIn using neutral job ads for delivery drivers, software engineers, and sales associates, and find a statistically significant skew in delivery by gender in Facebook's case for all three job categories.
@Techmeme @JeffHorwitz Thanks for writing! Here's the full paper https://ant.isi.edu/... and a thread with methodology details and implications: https://twitter.com/...
“How many times do researchers and journalists need to find these problems before we just accept that the whole ad-targeting system is bankrupt?” adds Christo Wilson https://twitter.com/...
“Though no audience was specified on the basis of demographic info, a feature Facebook disabled for housing, credit, & job ads in March 2019 after settling several lawsuits, algorithms still showed the ads to statistically distinct demographic groups.” https://www.technologyrevie…
Facebook disproportionately shows certain types of job ads to men and women, researchers have found, calling into question progress in rooting out bias in its algorithms https://www.wsj.com/... via @WSJ
Not enough has been written about how @LinkedIn seems to be doing better work on actually studying and addressing algorithmic bias issues within its own systems than the big companies that talk about it most. https://www.wsj.com/...
If you advertise a job opening on Facebook, Facebook will *automatically* deliver it disproportionately to people of the gender who already have that job. Is that just how advertising works in 2021? Nope! The researchers found that LinkedIn's job ads were delivered in a fair way …
“In tests run late last year, ads to recruit delivery drivers for Domino's Pizza Inc. were disproportionately shown to men, while women were more likely to receive notices in recruiting shoppers for grocery-delivery service Instacart Inc.” https://twitter.com/...
Despite its commitment to battle AI bias, Facebook's algorithms are still skewing the distribution of job ads according to traditional gender balance. The problems with race and age are likely the same. And there's no clear sense that's going to change. https://www.wsj.com/...