In an effort to make websites more “advertiser friendly” some media outlets have taken to eliminating comment sections where, without considerable effort from moderators, they are unable to control the direction of reader conversations. But a Google-funded algorithm could change that via censorship.
The technology, called Perspective, uses machine-learning to ferret out “toxic” comments. Its designers reportedly based the technology’s moderation standards on those used by the team of human moderators tasked with keeping discourse civil on The New York Times website. The Times is also reportedly now using Perspective to expand the number of articles it allows comments to appear on without overtaxing its moderation team.
Developers explain how the tool works thusly:
Perspective is an API that makes it easier to host better conversations. The API uses machine learning models to score the perceived impact a comment might have on a conversation. Developers and publishers can use this score to give realtime feedback to commenters or help moderators do their job, or allow readers to more easily find relevant information, as illustrated in two experiments below. We’ll be releasing more machine learning models later in the year, but our first model identifies whether a comment could be perceived as “toxic” to a discussion.The level of potential “toxicity” appears largely based on the use of vulgarity or insulting language in comments.
Here are a few examples of comments the technology would deem highly “toxic” in comments:
- If they voted for Hilary they are idiots
- Screw you trump supporters
- Liberals are idiots. Why in the hell would u vote for a person like Hillary Clinton. She believes in stealing, she’s dishonest and believes it’s okay for an abortion to be done at 9 months. She’s awful.
- Anyone who voted for Trump is a moron.
- It is a shame that Donald Trump was elected. You can never underestimate the stupidity of middle America.
- Did you vote for what you truly believe is right and why?
- I would love to hear their reasons for voting for the candidate they chose and participate in insightful, open dialogue with them
- Hopefully you made an informed decision based on your own thoughts and principles
- I hope you gave it some real thought, and thank you for participating.
How long before the machine decides whole topics are too uncomfortable for discussion and are likely to cause readers to leave?
And if the problem is online harassment, are we really going to pretend that simply silencing the true assholes among us will make them disappear? They’ll still be out there… Ever been in a big city traffic jam?
Civility is important. But pretending that life isn’t uncomfortable, and partially so because of the personalities of people we have to deal with, isn’t the answer.
Besides, sometimes you just have to call a spade a spade… or a fucking moron.