Increasingly, social media algorithms are revealing users’ sexual orientation or gender identity (SOGI) before they have consciously come out to themselves or others.
This is according to new research — “Contradictory connectivity and hybrid spaces: LGBTQ+ identity negotiation in the Hunter region of New South Wales, Australia” by Justin R. Ellis, et al — that was published in Gender, Place & Culture.
The researchers introduced a new concept – “algorithmic outing” — which is when digital platforms infer and surface a user’s LGBTQ+ identity through personalized content and recommendations based on engagement signals such as liking a video, following a creator or lingering on certain posts.
For this study, the researchers researchers conducted in-depth interviews with 20 LGBTQ+ adults aged 18 to 60 to explore how LGBTQ+ adults navigate “hybrid spaces” — or everyday environments where online activity and in-person interactions overlap, such as cafes, public transportation, workplaces and nightlife — that were seen as locations where “algorithmic outing” often occurs.
According to lead author Ellis, the phenomenon was becoming increasingly prominent as social media continued to advance. “Several participants told us the algorithm ‘knew’ they were queer before they did.”
The study’s participants noted that platforms like TikTok rapidly populating their feeds with LGBTQ+ content based on likes, follows or viewing habits.
“That experience was described as validating for some, but confronting or even frightening for others—particularly when it happened in public or semi-public settings,” Ellis said.
This, thereby, created “heightened vigilance” as algorithmic systems increasingly shaped what appeared on screens. Here, the participants reported using privacy settings, multiple accounts or passive browsing as a form of self-censorship in public spaces, thereby avoiding unwanted attention, harassment or being outed involuntarily.
“Algorithms sort, predict and classify users in ways that reflect existing biases and social tensions. For LGBTQ+ people, that can mean negotiating safety not just from other users, but from the technologies themselves,” Ellis said.
The research’s recommendations included particular design solutions, e.g. clearer consent mechanisms, privacy-by-design features, and tools allowing users to quickly exit sensitive content in unsafe situations.
