Technology adds automated transcripts, translation and identification of public figures

Reuters logo

Reuters has added AI-powered discoverability tools to its Reuters Connect software. 

The new technology aims to give users the ability to quickly find content through automated transcripts, translation and identification of public figures in videos. The Reuters innovation team built the features in house by fine tuning machine learning (ML) models to produce automated analysis.

Users will find time-coded, speech-to-text transcripts organized in interactive scene lists, as well as the ability to translate multiple languages into English. The technology also identifies and tags public figures within videos so that users can navigate to scenes that they feature in.

In addition, subscribers can replay live events, check soundbites and create clips on in-progress and concluded live events, allowing for instant broadcasting of Reuters content as well as streamlined publishing of highlights and social content. 

Separate from the AI technology, the service’s metadata has been enhanced.

Head of Reuters News Agency Sue Brooks said: “Our team of data scientists continuously research the latest AI technology to improve speed and efficiency for our customers. These new features harness the power of AI to save Reuters Connect users valuable time in getting Reuters content in front of their audiences.”