AI for Journalism & Podcasts: aureka Presents Latest Developments
Read Time 2 mins | Written by: Jan Kühn
aureka uses cutting-edge technologies to simplify the workflows involved in indexing and editing audio and video material. We are currently developing two tools that are designed to make work easier for journalists and podcasters in particular: A semantic search based on artificial intelligence (AI) and a production assistant to cut out parts from multimedia files with just a few clicks.
This development is funded by the Media Innovation Center (MIZ) Babelsberg, an institution of the public media authority Berlin Brandenburg (mabb). At the third "Innovation Pitch" on February 29, the aureka team presented the state of aureka's developments.
Searching for concepts instead of keywords
The audience was particularly interested in the new AI-supported search that aureka is currently developing. This allows users to search their collections of transcripts even better than before. In contrast to the existing full-text search, the semantic search works with concepts instead of just keywords.
AI is used to search for passages in the text that have a semantic similarity to the searched text. Semantics is a term used in linguistics and describes the meaning of words. For example, synonyms have a high semantic similarity - because the meaning is the same. Whether you search for "fast", "quick" or "nimble" makes little difference in this sense because the meaning of the words is very similar.
This makes it possible to find text passages that use different words but are similar in content. This function will become an integral part of the aureka Web App.
Audio quotes with just a few clicks
The second new feature will then improve the integration of the audio quotations found into existing workflows. With just a few clicks, the found passages can be cut out of the original multimedia files and reused. For example, long recordings can be searched for relevant passages so that individual excerpts can then be used in podcasts or other audio productions.
Thanks to these innovations as part of the "aureka for Media" project, aureka is establishing itself as a practical tool for anyone who works with audio or video material and needs to search it for relevant passages.
aureka can already transcribe and translate audio and videos automatically and generate subtitles in no time at all. Our AI also helps with the automated keyword tagging of content and it is possible to manually annotate the transcripts. The new functions take our app to a new level and make it an extremely practical research and production assistant.
Picture: © MIZ Babelsberg / Benjamin Maltry