Remember the IBM computer Watson that defeated humanity on Jeopardy not long ago? Last week, Watson came to mind again when I attended a compelling meeting of writers and programmers. It made me wonder how long till Watson and kin take over hunks of tomorrow’s newsrooms.
It started with a presentation by Andy Boyle on a program he’s writing called FireTracker. As I understand it, the program is designed to automate parts of the process associated with pulling together standard news stories on local fires. The idea is to make the generation of these kinds of stories more efficient and productive and, therefore, less expensive.
The idea is also to collect information in such as way that it becomes part of a database that can be explored and mined, one that can even allow reports on fires to be compared with one another over time with the hope of seeing patterns and gaining insights.
In his presentation, Boyle was assuming a journalist would still be on the scene, collecting or verifying facts, with a computer program helping to guide the parts of the process that can be most easily automated.
But it gave me flashbacks to how Watson was able to pick facts out of a giant database and respond to specific, arcane questions in fractions of moments. This made me wonder if a Watson heir will someday be continuously drawing data from public databases and then inserting information directly into certain story templates. I can envision a process that is mostly automated, assuming the Semantic Web technologies work as advertised.
I wish I could see this as an unvarnished good, the way I viewed most technology advances back the halcyon Clinton Era. Alas. I can easily imagine professional journalists, with their training in ethics and their accuracy-first values, being cut out of the game by start-up media companies that focus almost entirely on automated news and data visualizations. In such a paradigm, you need very few trained journalists, with most of the “analysis” written by bloggers, advertisers, consultants, and others.
Don’t get me wrong. That’s not my preferred scenario. There will be room for real journalists using the kinds of tools that Boyle and others are advocating. In fact, I don’t think journalists will have any choice but to embrace these technologies over the long haul. “Interactives” have already become serious players among major newspapers.
But we shouldn’t kid ourselves. Travel agents, bank tellers and automobile line workers weren’t just “complemented” and enhanced by automation and software. In many cases, they lost their jobs to it. They either reinvented themselves or became professionally extinct.
Now the same trends promise to affect skilled knowledge workers, from attorneys to journalists to medical diagnosticians.
Media organizations should keep an eye on this “automated journalist” trend, looking out 5 to 10 years to gauge how it may play out. They should formulate scenarios and potential strategies, working to ensure these technologies are used as wisely and ethically as possible. Because there’s no doubt in my mind that if journalists don’t think through these issues today, others will do the thinking for them in the not-so-distant future.