Los Angeles Times employs “robot reporter”


Technology has enabled cultures and societies to do the most work in as little time as possible. Before Google, we had to look things up in libraries, going through thousands upon thousands of pages of text and shuffling through drawers of catalogs. Before Facebook, we had to seek out friends and family through the phone book or through word of mouth. Before the wheel, we had to use our own two feet to get from point A to point B.

Now, technology has advanced to the point where I could write an algorithm that could possibly write this entire post for me, and no one would be the wiser. That’s what journalist Ken Schwencke is doing—instead of personally composing articles for publication, he developed a set of step-by-step instructions that can take a stream of data, compile the data into a pre-determined structure, then format it for publication. Schwencke’s algorithm works with earthquake statistics, since he lives in California.

Schwencke can go to bed at night knowing his work is being done overnight, and thus increase his productivity using the power of technology. He won’t have to spend late nights at the office or even look at a computer screen or touch a keyboard after he gets the algorithm running. He could then allot that time to doing what he personally wants while getting things done. “I doubt that people who read our (web) posts—unless they religiously read the earthquake posts and realize they almost universally follow the same pattern—would notice,” Schwencke said. “I don’t think most people are thinking that robots are writing the news.”

Algorithms are interesting pieces of technology—they are very versatile and can do many things. Auto-correct, the thing on your smartphones that may save or embarrass you when you’re texting your friends and family, uses an algorithm to figure out what you’re trying to say next. Remember those chess grand masters that played computers in chess? Those computers also use algorithms to determine their next move. “Algorithms can be highly complex computer codes or relatively simple mathematical formulas. They can even sometimes function as a recipe of sorts, or a set of repeatable steps, designed to perform a specific function. In this case, the algorithm functions to derive and compose coherent news stories from a stream of data,” said Jamie Dwyer, who provides IT support to Environment Canada.

“The use of algorithms on routine news tasks frees up professional reporters to make phone calls, do actual interviews, or dig through sophisticated reports and complex data, instead of compiling basic information such as dates, times and locations. It lightens the load for everybody involved,” Schwencke said. But, like many things involving technology, there are questions about the ethics of having an algorithm do all the work for you, and many journalists and academics have weighed in on the topic.

“Algorithms, like human beings, need to decide what is worth including, and make judgments on newsworthiness,” said Alfred Hermida, who teaches a course in social media at the University of British Columbia. “If the journalist has essentially built that algorithm with those values, then it is their work,” Hermida said. “All the editorial decisions were made by the reporter, but they were made by the reporter in an algorithm.”

“Many of the algorithms we encounter everyday exist in a black box of sorts, in which we see the results, but do not understand the process. Understanding how the algorithms work is really important to how we understand the information,” Hermida added.

Like earthquake statistics, sports statistics may be an area where something similar to Schwencke’s algorithm can thrive. “Baseball may be a good avenue for news algorithms, because the game is heavy with statistics,” said Paul Knox, an associate professor for the School of Journalism at Ryerson University in Toronto. “But even if an algorithm can analyze and manipulate data fairly well, journalism is still based on not only filtering, but also finding other available information, and a mathematical construct lacks the ability to dig up new facts or add context,” he notes.

Ben Welsh, a co-worker of Schwencke’s at the Times, also gave his opinion on the use of algorithms to write articles. “People are already reading automated data reports that come to them, and they don’t think anything of it,” he said. Welsh also mentioned smartphone apps that display personalized weather based on the owner’s location. “That’s a case where I don’t think anyone really blinks,” Welsh said. “It’s just a kind of natural computerization and personalization of a data report that had been done in a pretty standard way by newspapers for probably a century.”

Welsh, however, warned that algorithms should not completely replace the human journalist. “Responsibility for accuracy falls where it always has: with publications, and with individual journalists. The key thing is just to be honest and transparent with your readers, like always,” he said. “I think that whether you write the code that writes the news or you write it yourself, the rules are still the same. You need to respect your reader. You need to be transparent with them, you need to be as truthful as you can… all the fundamentals of journalism just remain the same.”

“Ultimately, it’s not about the tool,” said Lisa Taylor, a lawyer and journalist who teaches ethics at Ryerson University. “At (the algorithms’) very genesis, we have human judgment. Using algorithms ethically and reasonably shouldn’t be difficult; the onus is on the reporter to decide which tools to use and how to use them properly,” she continued.

“The complicating factor here is a deep suspicion journalists and news readers have that any technological advancement is going to be harnessed purely for its cost-cutting abilities,” Taylor added. “How can we use this effectively, reasonably, and in a way that honors the (tenets) of journalism?”

Source: Vancouver Sun