In many industries, cutbacks and layoffs have created more and more pressure on the remaining personnel to do more work. Longer hours, making do with fewer resources, and taking advantage of new technology to improve productivity is commonplace in many industries. Journalism and news reporting is no exception.

I recently read an interesting article about the rise of “automated reporting,” where computers are using algorithms to sort through data and create what passes for readable prose. In areas where there are vast amounts of raw data available, the “reporter” might just be a computer algorithm.

At Northwestern University, a group of journalism professors and computer science faculty have developed software to create machine-generated news content from data. This software has been licensed by the organization Narrative Sciences, an organization in the Chicago area that currently has several dozen customers of its machine-generated content business. It’s another area where artificial intelligence — the ability of computers to emulate humans — is making slow but steady progress.

Sports stories are an obvious application of the technology. First of all, if we consider all sporting events, not just professional, there are far too many to cover by human reporters.

Further, any sport today is ripe with data. With the help organizations that are focused on providing sports information, more and more data is being generated and gathered each day for every event.

Computer algorithms can sort through this data and automatically generate a basic story of what happened in the game. They may yet not be able to provide true insight into how the game was won or lost, or critical turning points, but the technology is demonstrating that it can provide something that passes for a game narrative.

There is a tremendous market for creating stories of local sports events. These are events that would not otherwise be covered if not for this technology. There simply aren’t the resources available to have reporters write stories for every event.

One example, so far: Narrative Sciences has produced over half a million accounts of Little League games, and that number is growing. The articles are generated using pitch-by-pitch game data that is actually entered by parents during the game using the iPhone application GameChanger.

Further, the game stories can be made available almost immediately after the game is complete. If Grandma and Grandpa don’t have a computer and access to the internet just yet, here is another incentive for them to get them.

There are of course many other applications. The New York Times has used this content-generation technology to create automated wedding announcements from reader-submitted data.

Earnings reports are another example. Forbes.com is exploring the approach for earnings reports and some financial news. The data is often there, it just needs to be cobbled together into a readable structure.

With some automated articles, not only is there no human author but sometimes there is also no human editing. Here there is certainly the potential for embarrassment. I haven’t yet heard of any major blunders yet, but let’s face it, the technology is still in it’s infancy.

There is an ongoing effort to fine-tune the algorithms to make the results better, or some may say “less bad.” However the Narrative Sciences articles I have seen seem very good, and in many respects might pass the “Turing test” for a readable article. That is, if you can’t tell that a machine wrote the article instead of a human reporter, then that might just be good enough.

Might computers be the reporters of the future? As much as I doubt it happening any time soon, if ever, I believe it certainly has the potential to help free up resources, human resources that can focus less on the mundane and more on the interesting, insightful angles and stories.

 

Leave a comment

Your email address will not be published. Required fields are marked *