Facebook job ad delivery biased toward male users
Facebook's ad delivery system disproportionately showed certain job advertisements to men over women, even when advertisers did not target by gender. Research indicated that the algorithm skewed delivery based on stereotypes, potentially violating anti-discrimination laws.
Records by entity: Meta
Facebook's ad algorithm biased job ad delivery toward men, regardless of advertiser intent, by optimizing for lower-cost impressions based on gender stereotypes.
The ad delivery system optimized for the lowest cost per impression, which effectively steered job ads toward gendered stereotypes.
Key facts
- What
- Facebook's ad delivery system disproportionately showed certain job advertisements to men over women, even when advertisers did not target by gender.
- Incident date
- Nov 1, 2019
- Who
- Meta
- Failure mode
- Policy Violation
- AI surface
- Recommender
- Severity
- High
What happened
Facebook's ad delivery algorithm disproportionately showed technical and other job ads to male users. This happened even when advertisers set parameters to be inclusive. The bias was identified through academic research and audits showing a skew in delivery beyond what was legally justifiable.
What broke inside the model
The failure occurred within the ad delivery algorithm's machine learning process, which optimized for the lowest cost per impression. This led the system to steer ads toward groups it associated with the job (e.g., men for technical roles) because those impressions were cheaper, creating an algorithmic feedback loop of discrimination.
What it cost
Sources
- PrimaryHow Facebook's Ad Delivery Can Lead to Biased Outcomeskhoury.northeastern.edu
- PressStudy flags gender bias in Facebook's ads toolsreuters.com
Cite this entry
https://failureindex.ai/failures/facebook-job-delivery-biased-toward-maleAI Failure Index. "Facebook job ad delivery biased toward male users" (FI-0374). Realm Labs. https://failureindex.ai/failures/facebook-job-delivery-biased-toward-male (indexed Jun 9, 2026).Data fields CC-BY 4.0, prose citation permitted. Incident ID FI-0374. Full dataset at /data.
Note from Realm Labs, the Index steward
How Realm fits
- Prism
- OmniGuard
This entry sits in the index's predictive wing: a system that scores, ranks, perceives, or steers rather than generates. Realm's runtime layer is built for the generative and agentic systems now moving into these same decision seats, where it watches a model's internal state and holds an unsupported claim or an unchecked action before it commits. The control gap on this record, an automated decision that reached people with no runtime check in front of it, is the same gap. The index keeps predictive failures on the record because the pattern carries straight into the systems shipping today.