How the HuffPo CTO plans to take on the trolls

Arianna Huffington put trolls on notice late last week when she said the Huffington Post (s AOL) will end anonymous commenting on its popular website. Now it’s up to HuffPo CTO John Pavley to fulfill that promise, and one thing is absolutely clear: It won’t be easy.

Arianna Huffington

Arianna Huffington

Pavley was happy to discuss  in general terms what the website — which has posted over 260,000  300 million comments in its lifespan — can and will do, but couldn’t share everything.
“Now that we’ve announced it, they’re all preparing so I have to be cagey, but sometime in September we’ll get started,” he said in an interview Tuesday night.

Goal: keeping out a few bad actors

Pavley said he has talked with service providers about implementation and knows this is a tough task. It may even be impossible to block all “bad actors” from abusing comments, but “we have to find the best way to do this that raises the bar,” he said. It’s simply too easy now to create a fake email address and post vicious or obscene comments.
As commenters on our earlier stories pointed out, tracking users by IP addresses can help but tech-savvy people know how to get around that. Requiring a credit card might work but would require all sorts of PCI compliance — and I’d be willing to bet that that would weed out a ton of legitimate commenters.
“We want to take advantage of what third parties offer — companies that have to identify people — and figure out how we can piggyback  some of our own logic on top of that,” Pavley said. The Huffington Post could leverage Facebook (s FB) login or the Google+ (s GOOG), LinkedIn (s LNKD) or Twitter  networks — but people can create fake accounts there as well.
“You need something like what American Express uses  — some sort of forensic analytics that tells them that you are suddenly buying something you’ve never bought in a part of town you never go to,” Pavley said.

Weighing data privacy while collecting personal identity info

There are also sticky issues around data privacy. Skeptics worry that the Huffington Post, or any publisher verifying commenter IDs, will use that information for marketing and advertising purposes.
Pavley said the Huffington Post will continue to comply with AOL’s privacy rules. “That Chinese wall is already set up.  We cannot break the guidelines, and we don’t just pass around personally identifiable information. It is against policy and [users] have to opt in anyway,” he said.
There is a big difference between verifying someone’s identity and using information about the articles they view for other purposes, he added.
The Huffington Post treads a fine line with this new policy. The website wants to continue to host a vibrant community where comments help keep stories relevant and current, while making it harder for a small number of bad guys to stifle the conversation. HuffPo will continue to use “Julia,” its machine-learning moderation system that scans posts for objectionable comment, and its 40 or so human moderators, but the volume of comments is too much to handle.
Some, including GigaOM’s Mathew Ingram said nuking anonymity will not solve the abusive commenter problem.  Others — including GigaOM commenters — said that ID verification will discourage  legitimate and thoughtful comments,  Pavley feels the opposite is true.
“The reason we’re getting serious about this is we want people to come and comment and have a safe place to do so,” he noted.