en.Wedoany.com Reported - South Korean live streaming platform SOOP (숲) has launched the AI admin service SARSA 2.0, allowing streams to continue even when the broadcaster is away. This AI not only assists with chat but also learns the broadcaster's tone and streaming style, directly engaging with users and recommending content.
SOOP recently released a beta version of the AI admin service "SARSA (SOOP AI Realtime Streaming Assistant) 2.0" to broadcasters and users. SARSA was initially a tool to help broadcasters efficiently manage their streams, featuring functions such as stream setup support, AI stream summaries, and chat atmosphere analysis. The core update in version 2.0 is the "stream takeover" function, where the AI takes the lead in the stream. When the broadcaster temporarily steps away or conducts an uninterrupted sleep stream, the AI can interact with users and recommend content. This version also includes a feature for co-watching existing video-on-demand (VOD) content to maintain a live stream atmosphere.
The main feature of SARSA 2.0 is that it reflects each broadcaster's "persona." By learning from the broadcaster's past VODs and stream data, the AI reproduces their unique tone, hosting style, and overall stream atmosphere. It can engage in real-time conversations based on user chat content and stream context, incorporating the broadcaster's preferred topics and content tendencies. The service operates in conjunction with the broadcaster-exclusive program "FreeShot Plus (프릭샷 플러스)." Broadcasters can choose to enable the AI admin in stream settings, manually activate the feature, or have the AI admin automatically take over the stream when the system detects the broadcaster has been away for a certain period.
SOOP has validated this feature in actual streams. Popular broadcaster "Murdoch" piloted SARSA 2.0 during a stream on the 6th of this month. When Murdoch stepped away, the "Murdoch SARSA" avatar interacted with users using Murdoch's own voice and tone, with viewers showing interest in the AI's natural expressiveness.
The industry views this attempt as the beginning of an AI broadcaster model. This model does not aim to completely replace human broadcasters with AI, but rather serves as a co-host to fill gaps when the broadcaster rests or is absent. Broadcasters reduce the physical burden of long streaming hours, while the platform maximizes user touchpoints by providing 24/7 uninterrupted content, creating a win-win structure. A SOOP official stated that the plan is to first apply SARSA 2.0 to a select group of broadcasters to gather data, then gradually expand the service scope, and continuously enhance the AI admin functions based on feedback from broadcasters and users, enabling its application in various streaming environments.
This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com
