Social media platforms constitute a space of both positive and negative gender dynamics. While the platforms challenge gender norms and empower individuals, they reflect and amplify existing societal biases.Therefore, we understand that we should not approach to social media as (inherently) sexist or not but try to recognize the interplay of factors underlined by the increasing use of AI systems(combining the user behavior, platform policies, algorithmic design and societal reflections).
We tried to test 3 different social media platforms to see how we would encounter with experiences of sexism by creating new profiles, one female account under the name of Emma and a male John, in order to eliminate the predetermined suggestions depending on our existing user history and test the influence of gender on the experience in these platforms.We tested three different social media platforms:
The results of individual tests were compatible with the generally accepted ones of the academic literature, however, the increasing improvement of social media platforms and their current state which differs from the problems of sexism faced within their creation, highly developed by research studies made it complicated to arrive at concrete and generalizing proofs of sexism and AI systems' implication.
The analysis of research studies and articles highlight the main idea that social media is a space of complexity and interplay of negative and positive aspects highlighted with the use of AI systems such as: Content Moderation, Personalization Algorithms, Ad Targeting and Recommendation, Trend Analysis, User Authentication and Security, Community Health and Sentiment Analysis.
Firstly, one of the most discussed concerns over sexism on social media platforms amplified with AI systems was the Representation imbalances: Through existing societal pressures and discriminations, women have been facing the problem of underrepresentation of their content and identities on social media platforms, however this problem is currently being amplified with the use of AI automated systems of recommendation and moderation of content as well as trend analysis. In this sense, men dominate discussions and hold more visible positions on platforms when considering the higher level of visibility and mentions, meanwhile the content of women face biased criticism and less consideration. (Journal Communication, Twitter retweets and mentions)
Similarily, another problem that the use of AI systems risk aggrevating is the practice of Stereotyping on social media platforms: The platforms perpetuate gender stereotypes through contents, advertisements and algorithms.They reinforce existing biases, lead to the amplification of stereotypical content with women more associated with aesthetically/ sexually pleasing content and domestic settings unlike men associated with professional and outdoor contexts, passing by AI systems such as personalized algorithms based on trend analysis and content moderation(Sex Roles, 2018, Instagram).
Furthermore, within the increasing significance and use of AI systems, the issue of Algorithmic bias has been developed: The AI algorithms used can inadvertently contribute to sexist outcomes especially when these automated systems are built in a way that disadvantages women: Depending on the design of systems and their learning process, AI systems such as Ad targeting and recommendation can end up displaying job ads depending on gender, or personalized algorithms can lead to a higher rate of error of facial recognition systems for women and people with darker skin tones (Nature 2018)
Last but not least, an important concern over the topic that touches the sensitive problem of security, is Online harassment: The actual effectiveness of User Authentication and Security and Community Health and Sentiment Analysis AI systems are open to a debate:Women frequently experience gender based harassment on social media, which means that online abuse (threats, derogatory comments, stalking) disproportionately targets women. (The European Institute for Gender Equality reported that women were more likely than men to experience cyber-harassment in all EU countries, with the highest rates observed among young women aged 18-29).
At the same time, AI systems and their implication on social media platforms can lead to a transformative collaboration, with important Positive aspects : Social media has empowered women to amplify their voices, connect with like-minded individuals, and engage in advocacy where AI systems such as trend analysis and content moderation play an important role. Studies demonstrate that social media has been a force for gender equality and provided a platform for marginalized voices (UNDP: advancement of gender equality by movements of momentum on social media platforms such as the #MeToo movement that can be related to AI systems such as trend analysis and content moderation).
