The use of Artificial Intelligence (AI) to replicate human intelligence is growing rapidly. New applications for AI keep being found, and current AI technology keeps being improved. I have a suggestion for online newspaper publishers: work with computer scientists to develop AI algorithms that can be more widely used to screen reader comments (see flow diagram).
I like browsing reader comments after I read a newspaper article or opinion piece. I almost always learn additional information, and reading the opinions of other readers helps me clarify my own views. I’m a “lurker” though not a commenter, meaning I seldom post comments myself. I got that term from a cousin’s husband. It seems apt.
For the five online newspapers I read regularly two allow comments and three don’t: Seattle Times (yes), Tacoma News Tribune (yes, on limited stories), Kitsap Sun (no), Oregonian/Oregon Live (no), San Diego Union-Tribune (no). The “no’s” have all changed their policies since I became a subscriber.
Even the Seattle Times has commenting restrictions. They don’t allow any when stories pertain to sensitive subjects. They also strive for respectful, civil discussion so hate speech, harassment, and profanity aren’t tolerated. That’s their call and it seems appropriate for a privately owned “family newspaper.” Unfortunately, commenting sometimes brings out the worst in readers particularly when they can hide behind an anonymous username.
So, it’s one thing to have a commenting policy and another to ensure users follow it by removing comments that don’t. Right now, for the most part, that takes humans. Humans need to be paid, and when they are moderating comments, even part time, they aren’t doing other news-gathering tasks. I suspect that’s why many online newspapers no longer allow comments. In a tradeoff between seeking a more engaged readership and staying profitable, profits win out.
There are smart people everywhere on the Internet with useful things to say, but on a newspaper website the readers have some common knowledge to start with assuming they closely read the article or opinion piece before commenting. That’s generally not true on social media websites like Nextdoor or Reddit no matter how carefully a topic is introduced. Thus, there can be an increase in value vs. time spent between reading newspaper comments and social media comments.
Back to using AI to screen newspaper comments. Could it really be done with an approach that readers would tolerate without thinking their creativity was being stifled or their innate mistrust of machines replacing humans kicking in? I think the process could be made to work, although it would have growing pains. If so, there’s an irony here. Comments reflect what people think about the article or subject, yet a machine would need to approve them. Why not just use AI to write the reader comments instead of moderating them? Just kidding on that one.
I’m tossing this out there for online newspaper publishers and computer scientists to work out the details. But if more newspapers enable reader comments again, I might need to reduce the number I read.
Mud Bay Blog 