Narrative Science raises $10M, taking it a step closer to automating this post

Narrative Science, the company that has developed software to automate writing reports based on structured data, has raised an additional $10 million in funding. The round, which brings overall funding to $32 million, was led by insurance and financial company USAA, and from existing investors Sapphire Ventures (formerly SAP Ventures), Jump Capital, and Battery Ventures. According to the press release that quotes Narrative Sciences CEO Stuart Frankel, USAA has been applying Narrative Science technology to “turn mountains of financial data into information that can be easily understood and acted on by millions of people.”

USAA is using the product to create internal reports.

Other customers include Forbes (see this earnings announcement page, powered by Narrative Science) and the CIA (see Peter Kafka’s CIA Venture Arm In-Q-Tel Invests in Narrative Science).

The company got its start when two of its founders, Kris Hammond and Larry Birnbaum, created Stats Monkey at Northwestern University, an application that could create passable sports news stories based on the box scores.

Now, if stories like this fundraising were accompanied by structured data — funding amount, companies participating, quotes from CEO and VC, etc. — they might be written in the near future by Narrative Science or its competitors.

But wait. Why can’t AI be pointed at the easily interpreted structure of press releases? If TripIt can parse emails with flight and hotel itineraries and turn those into calendar events on our behalf, surely it wouldn’t take much neural network tweaking to read PR and push out a bland but perfectly understandable story.

Another example of how algorithms and AI are digging into the work that well-trained professionals formerly did, and which the machines now do, is that of expertise in many realms being commoditized. Oxford researchers Carl Benedikt Frey and Michael A. Osborne, in a much discussed paper last year (see How many of today’s jobs could be computerized? A whole lot.), showed a detailed analysis of today’s occupations and suggested that as much as 47 percent of employees are at high risk of replaced by AI and robots.

Below are the jobs that are at low, medium, and high risk:

Screen Shot 2014-01-22 at 11.12.29 AM


The Narrative Sciences announcement is just another baby step in this direction, where perhaps as many as 400 million people will have to find other work as their occupations are taken over by AI.

Yesterday I summarized recent discussion on this issue by Thomas Davenport, Kevin Kelly, Nick Bilton, Nick Bostrum, Elon Musk, and Stephen Hawking. As I wrote in that piece:

This is clearly a major debate in society, and even if the result isn’t a shiny, shiny utopia as envisioned by Kelly, or the burnt earth dystopia that Bilton and Bostrum worry about, we are still likely to be stuck in the middle: living in a world where near- or past-human-grade intelligence can be rented for pennies on the dollar.

Every announcement like Narrative Science’s pulls us a bit closer to a day of reckoning, where we will have to decide what to do with the surplus workers that are pushed out of the market for work.