In the study, men could get about a meter closer to birds than women could before the animals flew away, according to the results. This pattern remained regardless of what the men and women were wearing, what their height was or how they tried to approach the creatures. That suggests birds may be able to suss out the sex of a human, though the researchers aren’t sure how.
...
In the paper, the team hypothesized that birds may be sensing chemical signals, such as pheromones, or using cues such as body shape to recognize a person’s sex. But more research is needed before they can come to any conclusions. Notably, previous findings in mammals also suggest these animals can tell men and women apart: for example, lab rats have been observed to feel greater stress when male researchers handle them than when female researchers do so.

“We have identified a phenomenon, but we really don’t know why. However, what our results do highlight is the birds’ sophisticated ability to evaluate their environment,” said study co-author Federico Morelli, an associate professor at the University of Turin, in the same statement.

Flight initiation distance (FID) is a metric often used to study an individual's perceptions of risk when facing a predatory threat. Longer FID indicates lower risk‐taking, while shorter FID identifies bolder individuals who tolerate greater risk. Until now, no studies have tested the potential effect of the observer's sex on the escape behaviour of wild birds. Given observed differences in how laboratory animals may respond to the sex of humans interacting with them, the lack of reports in the field is surprising. In five European countries, we tested whether urban birds perceived the risk posed by approaching female versus male observers differently, using FID as a response variable.

First, we matched the female and male observers according to their height and clothing. Then, we fitted Bayesian regression models, controlling for the phylogenetic relatedness of bird species, to test for the effect of human observer sex after controlling for a variety of other important factors known to explain variation in FID (starting distance, flock size, sex of the target bird, land use characteristics and vegetation cover).

We found that male birds were more risk‐tolerant than females and – unexpectedly—birds in general escaped sooner when approached by women than by men. The escape difference associated with the observer's sex (~1 m longer when approached by women than by men) was consistent in populations across all five examined European countries. We discussed various hypotheses to explain birds' escape responses related to the observer's sex; however, further research is necessary to fully understand this phenomenon