A brief history
The Deepnews.ai project started during the academic year 2016–2017 when Frederic Filloux was a John S. Knight Journalism fellow at Stanford University.
At the time, the “News Quality Scoring” project was based on the approach of what kind of “signals” could be drawn from a news story that would differentiate a value-added article from a commodity news piece?
The following year, the JSK program awarded Frederic a research fellowship to pursue this journalistic exploration. By that point, the project had become 100 percent artificial-intelligence driven, thanks to the advice of Dennis Allison, a computer science professor at Stanford. Hence the name Deepnews.ai.
In Summer 2018, when Frederic’s research fellowship ended, the project was moved back to Paris. Two engineers from Ecole Centrale de Paris helped lead the development of about 100 versions of the deep learning model — the Deepnews Scoring Model — as explained here.
In early 2019, it appeared that Deepnews.ai needed an experienced executive to bring the company to the next level, which would mean developing a lineup of automated newsletters powered by the Deepnews Scoring Model and growing the business. David Finch joined the company in March and will become the CEO in 2020.
In the summer Deepnews.ai launched the first demonstrator of what its technology can do, the Deepnews Digest. The weekly newsletter shows how there is good reporting happening out there, and that the Deepnews Scoring Model can find it.
This year Deepnews.ai expanded on the Digest by demonstrating that it can monitor subjects over time with several months’ production of Deepnews newsletters on various topics around tech, coronavirus and regions of the world. In parallel work continues on refining the model and building a robust infrastructure to enable rapid and reliable scoring of 1000s articles per hour.
By the end of the year a commercial API is made available to enable third parties to integrate Deepnews.ai into their projects and solutions. To fully focus on developing the core technology and serving B2B clients, the newsletters – with the exception of the Digest – are shelved.
Deepnews.ai has ambitious plans to develop a new version of the Deepnews Scoring Model with ability to score a far wider range of content as well as articles in languages other than English.