Decisions to restrict data access to major platforms like Twitter and Reddit have recently led to scholarly discussions of a “post-API” world, in which researchers can no longer depend on company-sanctioned access to powerful platforms (Bruns, 2019; Freelon, 2018; Tromble, 2021). Yet TikTok appeared to counter the trend with the release of a Research API in 2023. The API was designed to aid researchers studying TikTok, reducing the need to rely on external open-source tools (e.g. PykTok), where the scope of data may be limited.

Scholars granted access to the API can search for videos and users, and retrieve metadata on viewership, engagement with, and circulation of videos. TikTok allows researchers to retrieve up to 100,000 records per day. Such data could help provide insight into patterns of communication on one of the world’s fastest growing social media platforms, and one that is increasingly used by young people for social interaction, information seeking, and political speech (Cervi et al., 2023; Schellewald, 2021; Song et al., 2021).

Following the tradition of previous digital data quality audits (Pfeffer et al., 2023; Tromble et al., 2017), we perform an audit of the TikTok Research API, with an interest in helping researchers evaluate the utility and quality of the API. We conclude that research based on this data source may lack scientific validity, particularly if it relies on video metadata (likes, shares, views, comments). However, we also discuss potential uses of the Research API for archival searches.

In this paper, we explore how AI researchers, situated within university-based research networks, mobilise and resist industry interests. The research question to which this paper is addressed is: how do university-based academics in the field of AI experience and mediate industry influence in their research? We answer this question through semi-structured interviews with research-focused academics (n = 90) affiliated with university-based AI research networks. We find that national research funders and university leadership incentivise and facilitate industry investments in AI research. We demonstrate how AI researchers mobilise this interest to pursue their own research goals, whilst also—at times—subordinating their research goals to the interests of industry. We highlight how AI researchers internalise the commercial logics of technology firms, which become mirrored in researchers' orientation towards generalisable and scalable research outputs that can move between many application domains and local contexts. We argue that university-based AI research networks primarily operate as mediators between industry, government, and university actors, and highlight the role national research investment strategies play in creating an enabling environment for industry influence of AI research.

This paper charts the emergence of a distinct category of research-dedicated GenAI platforms, which we term Research GenAI or RGAI. These platforms are explicitly marketed to a cross-disciplinary academic audience, promising to automate research discovery and writing tasks, such as identifying/summarising published research, writing literature reviews, conducting data analysis, and synthesising findings. RGAI platforms (e.g., Consensus, Elicit, Research Rabbit, Scholarcy, Scite, SciSpace) are rapidly being adopted, in a context of experimentation, uncertainty, and controversy.

We define the contours of Research GenAI by mapping the history and development of RGAI platforms and developing a preliminary typology of RGAI. We situate RGAI platforms within the scholarly economy and ongoing processes of platformisation and automation of academic work. We make a case for the need to understand RGAI platforms as complex sociotechnical systems that intersect with social, ethical, institutional, and legal questions, and demonstrate this approach through an STS-informed walkthrough of two notable RGAI platforms: Consensus and Elicit. In this presentation we present our findings generated from these walkthroughs and explore the implications of the technologies for the academic publishing industry.

Companies that collect personal data have spent billions of dollars complying with a patchwork of global data privacy laws since 2018. In response, a nascent privacy tech industry has emerged, consisting of tech startups, consultants, investors, platforms, and domain experts that collectively help companies build compliant data governance programs. This study recognizes the key role of translating the law into software products by asking: how is expertise defined and encoded in the privacy tech industry? I draw on fieldwork from a broader ethnographic study of the privacy tech industry to identify three findings. First, the privacy tech industry constitutes a networked arena of relations structured by partitioning professional expertise across technical, legal, and operational domains. Second, technical expertise in the privacy tech industry is often tenuous and contingent, which could be strengthened by applying scrutiny and deliberation to evaluate the content of expertise rather than its performance. Third, boundaries of expertise are increasingly encoded in compliance software, perpetuating performative rather than scrutinized expertise. At scale, these products promote managerial processes that manifest as checkbox compliance, codifying splintered accountability and undermining the spirit of data privacy law. I argue that technology policy would benefit from scrutinizing the contents of expertise, inviting agonistic deliberation, and including lay expertise to counter the technocratic structural relations of a surveillance economy.