If you read almost anything online, there is a good chance that you have read something written by a machine without even realising it. From financial reports to football commentary, machines can write quicker and more cheaply than any journalist and in many cases they can also write more accurately – machines do not make typos and rarely make copying errors, though they are prone to much more serious problems.
I have written about this before (www.marinetraffic.com/blog/ais-impact-maritime-crisis-communications) and made the argument that the role of AI in the newsroom was likely to focus on supporting human journalists by doing those tasks that computers are especially good at (e.g. identifying patterns in large data sets) rather than trying to do things that humans are usually much better at (e.g. understanding human emotions). It appears I may have been wrong.
If there is one thing people are better at than machines surely it would be being a person, but China’s official state-run news agency, Xinhua, recently debuted an artificial news anchor which they claim "can read texts as naturally as a professional news anchor” – a significant exaggeration in my opinion, but still a potentially significant development in the role that computer programs will play in how we consume news.
If programs can research, write and now even present the news by themselves, do we still need journalists?
Yes, and perhaps more than ever.
Too many news outlets, facing tightening budgets and the temptation of advertisement revenue which is maximised by attracting as many clicks as possible, have focused on quantity over quality – finding that as a business it is better to write 10 stories quickly rather than one really good story.
This has had some advantages for the public relations industry as more and more publications will publish word-for-word almost any press release they are sent, but it has also made it that much more difficult to manage rumours and baseless speculation.
Automating the process of “writing” stories is a natural continuation of this process. Xinhua can now produce an endless stream of breaking news potentially with very little (if any) human involvement. This power will help them to dominate the news feeds of countless people.
However, “churnalism” – simply repeating and recombining what is already out there on social media, blogs, other media sites, in company press releases and so on – should not be confused with journalism. A real human journalist has the ability to delve a little deeper, ask questions and most critically to sense check the information being reported. Does it always work? No, but I think we should at least demand a human check and balance on the information we use to make decisions.
As AIs and human interfaces get increasingly clever journalists must increasingly focus on what they are best at – critically investigating a story, asking hard-hitting questions and, dare I say it, trusting their gut. If journalists allow themselves to simply repeat what they read on social media and pick up from other easily available sources, then they will soon be replaceable, but if they research and develop their stories, then they will remain indispensible for at least as far into the future as I can see.