Saturday, February 21, 2015

Computers as Journalists? NO, thanks.

There's this company, Narrative Science, and they specialize in creating computer generated narratives, articles, etc. With the help of the Northwestern University school of Journalism, they have been having computers write data-driven articles for Forbes and other media outlets. Whaaaaaaaaat?! A computer writing articles? This is crazy. But apparently, you can hardly tell when you read it. The Chief Product Officer at Forbes, Lewis DVorkin, says that it's actually been helpful to the journalists because they can focus on other things that they would rather write and let the computers do the other,  more data focused stuff. He says they attach the story arcs to the data, and they merge, creating a well written story.
The company has perfected the algorithms for finance and sports score stories, but here's the thing. What about human interest stories? A computer can be super smart and organize data into a written story, which is actually pretty cool. But a computer can't pick up on human interaction and be able to tell a story as eloquently as a writer, you know, someone who dedicated their life to observing humans and documenting it. I just don't buy it. But according to the co-founder of Narrative Science, Kristian Hammond, computers are projected to write 90% of news in the next 15 years. So people are freaking out because that means no more journalist jobs.
As gatekeepers to the media, human journalists can sometimes fail in being completely objective. They shape the way we see the world, and what's going on in it, but it's impossible to write about everything, so not all the important things are covered. But news outlets strategically plan what they will and won't cover, and how. I guess the computers could be helpful, in covering sections of the news, because humans simply can't do everything, and users can even customize the tone of the articles, from "breathless financial reporter to dry analyst". Narrative Science has also figured out that it could be helpful in analyzing huge data information. So if you are trying to look at hundreds of pages of spreadsheets, the computer can go through it and summarize it for you in a written paragraph. That's pretty cool. But as far as writing journalistic stories, I don't like it, even if you can customize it.
The algorithm learns from other published articles, and figures out the important aspects to include. But I worry that news stories will become cookie cutter, with no individuality, so that everything we read will be pretty much the exact same format and tone. Humans integrate their individual tone and sometimes personality in their writing, and it's enjoyable. Computers don't have experiences, and experiences are what shape us as humans. I also worry that, since the computers look at other articles and learn, they will continue to perpetuate the same issues with the media that we already have, rather than learning from it and changing it, because it learns and bases itself off of it instead. So since the media disproportionately shows Brown and Black faces when talking about welfare, crime, and poverty, wouldn't the computer do the same? I think we need more human journalists that are aware of these issues and doing what they can to change it, rather than a computer continuing to perpetuate prevalent stereotypes in the media.

2 comments:

  1. I like that take you chose to take on this past week's subject. It was refreshing to read about something that is present and so controversial. It is interesting how the world is becoming so reliant on technology. thanks for sharing!

    ReplyDelete
  2. Really interesting topic! I wonder if there is a positive to these cookie cutter news stories? Maybe they could start as a jumping off point for journalist and help to focus our news and weed out biased opinions.

    ReplyDelete