In my graduate school research I intend to study the impacts of newsroom automation and artificial intelligence on journalism. As part of that I am creating a few code experiments that explore some of the possible ways journalism will impacted by emerging technology.

It is believed that automation and AI will not eliminate the need for experienced journalists, but automation could still have a dramatic impact on newsroom operations as a whole. Certain newsroom rolls are more likely to being replaced by automation than others, and one particular vulnerable position is that of a news reader or presenter.

Significant advancements in the quality of text to speech technologies, has made automated audio based news presenting much more feasible. Advancements in text-to-speech technology have made it much more difficult to tell the difference between human and computer generated speech.

This site uses Python to take the publically available RSS feeds of a news organization (in this case the CBC) and automatically generate an audio news bulletin based upon a certain set of content criteria. (Currently 5 top stories, 2 business stories, 2 sports stories and 2 arts stories.) The text of the news bulletins are generated into speech using the Google Cloud Text-to-Speech API. This page is updated with a new “newscast” every 30 minutes.

Automated news presenting also opens the door for a newscast that is customized to the interests of an individual user. Whether this is a good thing is up for a debate, and is an issue I hope to explore further.

This project is still very much a work in progress. There are few changes that I would like to make that would make the newcasts sound a bit more natural, I would clean up the presentation of the player a bit more perhaps providing more context to the stories that are being heard and if I get ambitious allow users to customize the mix of content they would hear.