To meet the demands of content moderation, online platforms have resorted to automated systems. Newer forms of real-time engagement($\textit{e.g.}$, users commenting on live streams) on platforms like Twitch exert additional pressures on the latency expected of such moderation systems. Despite their prevalence, relatively little is known about the effectiveness of these systems. In this paper, we conduct an audit of Twitch's automated moderation tool ($\texttt{AutoMod}$) to investigate its effectiveness in flagging hateful content. For our audit, we create streaming accounts to act as siloed test beds, and interface with the live chat using Twitch's APIs to send over $107,000$ comments collated from $4$ datasets. We measure $\texttt{AutoMod}$'s accuracy in flagging blatantly hateful content containing misogyny, racism, ableism and homophobia. Our experiments reveal that a large fraction of hateful messages, up to $94\%$ on some datasets, $\textit{bypass moderation}$. Contextual addition of slurs to these messages results in $100\%$ removal, revealing $\texttt{AutoMod}$'s reliance on slurs as a moderation signal. We also find that contrary to Twitch's community guidelines, $\texttt{AutoMod}$ blocks up to $89.5\%$ of benign examples that use sensitive words in pedagogical or empowering contexts. Overall, our audit points to large gaps in $\texttt{AutoMod}$'s capabilities and underscores the importance for such systems to understand context effectively.

Surprise, Discord is partnering with OpenAI to integrate ChatGPT throughout the app. There's a chatbot, obviously, but the company also plans to use machine learning in a handful of more novel and potentially useful ways. Starting next week, the company will begin rolling out a public experiment that will augment Clyde, the built-in bot Discord employs to notify users of errors and respond to their slash commands, with conversational capabilities. Judging from the demo it showed off, Discord envisions people turning to Clyde for information they would have obtained from Google in the past. For instance, you might ask the chatbot for the local time in the place where someone on your server lives to decide if it would be appropriate to message them.

Discord is adding AI to its platform in the form of ChatGPT and generative art, which will manifest as a chatbot and options to manage chats and create custom avatar profiles. Discord plans to roll out a public ChatGPT-powered chatbot named "Clyde" beginning next week, alongside a new technology to summarize Discord chats in a sidebar, called conversation summaries. This Friday, Discord will update its AutoMod automatic moderation bot to include AI-powered moderation, examining the content of moderated chats to determine if a server's rules are being followed. All three are considered public experiments, with updated, further rollouts to come later. Discord also showed off early progress in two new features it hopes to add later: the ability to "remix" Discord avatars, as well as an updated real-time whiteboard feature that can take sketches and transform them into generative AI art, via a prompt.