arXiv Pilots AI-Powered Audio Summaries in Partnership with ScienceCast 

From the arXiv Blog:

Audio summaries are coming to arXiv! arXiv has partnered with ScienceCast on a pilot project to bring 60-second AI-generated audio summaries to arxiv.org. This follows our efforts to make arXiv’s scientific content even more broadly available by supporting additional formats beyond PDF, starting with HTML versions of recent papers, which we began rolling out in 2024. The pilot project with ScienceCast is initially limited to a small subset of arXiv papers, namely those in the astro-ph.HE category. However, we expect that it will be rolled out to more categories, and that other forms of audio content will be supported in the near future.

ScienceCast has a long-standing relationship with arXiv, and has successfully partnered on a similar pilot with bioRxiv since November 2023. Their AI-driven platform allows authors to create audio and visual enhancements for their scientific papers. They became an arXivLabs collaborator in 2022, making ScienceCast content discoverable on arXiv through the “Code, Data, Media” tab on the abstract page. This new project with ScienceCast automatically generates 60-second audio summaries for select arXiv preprints.

[Clip]

Audio summaries are currently only available in high energy astrophysics (astro-ph.HE) for new papers, produced from TeX, posted to arXiv on or after December 1, 2024. Audio summaries are automatically generated shortly after each new paper is posted to arXiv, and a link to them will be available in a menu on the paper’s abstract page.

Learn More, Read the Complete Announcement

Filed under: Data Files, Journal Articles, News