Influence magazine has made the case that machine journalism can increase productivity and audience relevance. Is this the future?
An article in Influence (Q4 2018) explains how access to open data sources in the UK, combined with natural language generation (NLG), has almost automated the production of media stories. The article says that the UK’s official published data offers good segmentation. For example, crime statistics are shown for each local authority area. This means that local stories about crime levels are relevant and specific to each town or region. The author argues that this flips the normal news agency business model “on its head.” Instead of providing central one-size-fits-all content, as some news agencies do, this model produces customized content targeted to a smaller segment of readers. This makes it ideal for local news outlets. Using NLG tools or “machine journalism” as some people call it, many stories can be produced fast by a few journalists. And each has great audience relevance because the data can be easily segmented. Local media always need relevant local content but often have too few resources to produce it. I was impressed with this concept and recommend the article. It could be the future.
Concerns about machine journalism
however, many people in PR and journalism have expressed concerns about this. And this is at the same time as acknowledging that natural language generation techniques increase productivity, as the article explains. But many people feel that a human is needed to add local context and personal case studies to give an authentic voice to each article. Readers are savvy and they can spot anything that seems de-humanised or automated. And if they see an almost identical article in a neighbouring local media news outlet, they might not be impressed.
Is this the future?
So whilst I welcome any innovation that supports the UK’s fantastic local media sector, I think people in PR and journalism are right to worry about this. We want our readers to engage fully with our articles and news in a real human way. We want to shape important debates, change opinions and explode myths. Is this really possible with NLG templates, even with the addition of verifiable public open data sources?