2024
You can’t say that on TikTok : cxnsxrshxp, algorithmic (in)visibility, and the threat of representation
[link]
Sydney Dawson (MA)
Abstract
In our current age of the internet, artificial intelligence, machine learning, and algorithms are a nearly ubiquitous feature of online experience. Social media is a key element of the diverse landscape of the internet, and the increased reliance on algorithms by these platforms in recent years marks a shift away from networking the self to algorithmizing the self, or understanding oneself not through social connections but through the versions of the self represented by recommender algorithms. TikTok, a short-form video content platform, has quickly become one of the most popular social media platforms in North America, partially because of its highly engaging content suggestion algorithm. However, algorithms are not neutral actors, and both reflect and shape the biases of developers, users, and other key stakeholders. Amidst tensions on social media about freedom of expression, and re-emerging moral panics over issues such as good citizenship, sex(uality), and the innocence of children, TikTok represents a unique and contested site of expression; users who prefer authentic and emotional engagement on social media prefer to use the platform to discuss issues which face social stigma, and other users feel that these expressions pose a danger to the citizenship and safety of TikTok as an online environment. As a result, malicious content reporting has led the platform’s moderation algorithm to enforce sets of biases and ideologies as pseudo-community guidelines, and users at odds with these new moderation measures are inventing creative ways to tactically (in)visiblize themselves through language, sharing messages with specific audiences while using the tools of algorithmic promotion to curate their audiences. This thesis identifies several linguistic categories employed to evade algorithms in playful, creative, and furtive ways, and takes up the roles of fear, danger, and threat in affectual responses that drive malicious content reporting. Through a gender and sexuality analysis, we can understand expressions of sex(uality) on TikTok and the efforts of users to invisibilize and police this content as efforts to correct and eliminate the sexually non-normative behaviour of others as a representation of visibility and acceptability politics and moral panics about queerness and sex in North America.
We’re cooked : creative responses to generative AI and atmospheres of technocapitalism on Twitter
[link]
Trinity Stewart (MA)
Abstract
Leading up to November 2022, generative AI art tools such as OpenAI’s DALL-E (January 2021), Hugging Face’s Craiyon (April 2022), and Midjourney (July 2022), were steadily gaining popularity. However, the release of OpenAI’s AI text-based tool, ChatGPT, in November 2022 brought many new eyes to the world of generative AI, boosting both the visibility, and eventually the prevalence of such tools. With increased attention from non-technical communities, generative AI became a topic of intense scrutiny online, with discussions and debates growing across many social media sites. One particularly vocal group has been Twitter’s creative community, composed of artists, voice actors, video game developers, and more, who have taken to the platform to share their diverse perspectives and concerns. This thesis explores the affective atmospheres that arose from Twitter’s creative community between November 2022 and February 2024, locating negative emotional responses to generative AI within the larger context of neoliberal and technocapitalist ideologies. I also examine Twitter as a place with agency, recognizing the interplay between users of the platform and its algorithm; this project is a qualitative analysis that relies on a set of tweets, made by creatives, and selected by the algorithm through interactions with Twitter’s “For You” page. In this way, AI tools are both a subject, as a key “villain” of the narrative, and through the algorithm, a collaborator. I close by moving beyond emotional responses to generative AI, focusing instead on efforts made by creatives to mobilize and resist exploitation.