YouTube recommendations – A study using automated social bots found YouTube’s recommendation system can steer male- and female-coded accounts into markedly different political information environments—even when both start with the same baseline interest in News & Politics.
For years, campaigns have understood the seduction of a screen. In the 2020 election cycle, the Trump campaign bought out the masthead ad space at the top of YouTube 20 times, including an audacious buyout on Election Day.
But the influence may not stop at ads. A new study published in the Cornell University repository arXiv argues that YouTube’s recommendation system can do something more subtle: actively sort male and female users into vastly different political information environments. even when their initial interests are identical.
The researchers ran experiments with 160 automated social bots—80 programmed with what they described as male-coded viewing habits such as sports and gaming. and 80 programmed with female-coded habits such as style and vlogs. Each bot was then given the exact same baseline interest in YouTube’s News & Politics category.
To test where the algorithm would take them, the bots completed 150 consecutive interaction steps. The point was not what the bots liked on day one, but what the recommendation engine pushed them toward as those interests accumulated.
The results were striking in different directions. Female-coded accounts encountered a higher overall volume of political videos. Yet the issues recommended diverged sharply by whether a profile carried male- or female-coded viewing habits.
Male-coded profiles were disproportionately funneled toward a narrow set of confrontational domestic issues, including law, crime, and defense. They were also pushed heavily toward state-power entities such as Immigration and Customs Enforcement and the Department of Justice.
Female-coded accounts, by contrast, were presented with a broader, more moderate mix of macroeconomic and lifestyle-related public policy topics. The study lists international affairs, culture, and the arts among the areas female-coded profiles were recommended. It also reports that female-coded profiles received significantly more neutral political content, while male-coded profiles were shown more polarizing videos.
The pattern didn’t just show up in topic selection—it also appeared in the shape of what the study describes as the information network. Male-coded profiles were trapped in a concentrated web of overlapping videos. repeatedly encountering the same content in what the study characterizes as a cohesive echo chamber. Female-coded profiles, the researchers found, experienced a far more diffuse and differentiated information network.
Jonathan Gray, codirector of the Center for Digital Culture at King’s College London, reviewed the findings but was not involved in the study. Gray said, “YouTube is one of the most widely used platforms on the planet, yet its algorithms remain opaque and poorly understood.”
Gray also pointed to the stakes of that opacity as platforms increasingly amplify harsher material. “For many it is a primary source for news, advice, and guidance,” he said. “In a moment where platforms are promoting increasingly misogynistic and extremist content. this study contributes to a growing body of work investigating the role that their algorithms play in shaping society. culture. and politics. highlighting an urgent need for greater public scrutiny and oversight.”.
The study also comes with a limitation that hangs over any conclusion: the authors did not respond to an interview request. Among the questions that would have been asked was how reliable it is to stereotype viewing habits using male-coded and female-coded labels like sports and gaming on one hand. and style and vlogs on the other.
YouTube recommendation algorithm political information Cornell arXiv study automated bots echo chamber gendered content Immigration and Customs Enforcement Department of Justice misogynistic content digital culture