Abstract
The internet, as a public space, is characterized by one important feature: anonymity. This anonymity allows users to access and engage with the online sphere while their individual identities and backgrounds remain relatively obscured, with the option to selectively disclose such information. Anonymity enables individuals to express their opinions and emotions more freely in public spaces, fostering diversity and representativeness in public discourse. It also shifts the focus to the content of discussions rather than differences in identity, thereby reducing bias and discrimination and promoting social inclusivity. Ensuring freedom of expression in the online sphere requires safeguarding this anonymity, especially in today’s era of political polarization.
However, some governments may eliminate anonymity to monitor and control information. For instance, in 2022, the Chinese government mandated that internet service providers display detailed user information on public account pages, including the operator’s name, social credit code, valid contact information, and the user's Internet Protocol (IP) address location. Given the severe regional disparities in China’s development, such disclosures, particularly users’ geographic locations, significantly undermine anonymity. This loss of anonymity fosters assumptions about users’ identities, which, when combined with biases against certain identities, can erode trust among the public and alter patterns of participation in public discussions. In this paper, we examine the consequences of removing anonymity in online spaces.
We conducted a randomized controlled trial with over 2,000 internet users to examine how identity disclosure impacts public perceptions and online discourse. Participants were exposed to both balanced and unbalanced conditions. In the balanced condition, they viewed TikTok screenshots simulating one-on-one debates, with speakers’ IP locations edited to test the effect of identity disclosure. In the unbalanced condition, additional comments created group attack scenarios targeting a specific user. Participants’ responses and prosocial tendencies were measured using variations of the dictator and ultimatum games. To ensure data accuracy, we verified self-reported IP locations against platform data, embedded trap questions, and set minimum response times for each question.