Toxic Comments

We are fortunate here at AVC. We have mostly civil and respectful conversations. People behave themselves here. That is sadly not the case everywhere.

I don’t know what the people who post comments like this are feeling and thinking. It is horrible. Awful. Hateful. Hurtful. Painful. Disgusting. Disturbing. And a lot more.

If you operate a large social media service like Twitter, Facebook, or Disqus, you get to see stuff like this every day, hundreds of times a day. It is a view of humanity that is deeply upsetting.

Disqus, which is a USV portfolio company, where I serve on the Board, and which operates the comment service here at AVC and at millions of other websites around the globe, has been working on scaleable solutions to this problem.

They posted an update yesterday on what they are doing to combat this problem.

Here are some excerpts from that post:

The Disqus Platform supports a diversity of websites and discussions; with such a large network of publishers and commenters, having a policy against hateful, toxic content is critical. While we do periodically remove toxic communities that consistently violate our Terms and Policies, we know that this alone is not a solution to toxicity. Oftentimes these communities simply shift to another platform. Ultimately, this does not result in higher quality discussions, and it does not stop the hate. In order to have a real, lasting impact, we need to make improvements to our product. Which is why, if at all possible, we work with publishers to encourage discourse (even unpopular or controversial discourse!) while helping to eliminate toxic language, harassment, and hate.

Over the past several months, many passionate folks have reached out to us about severe violations of our Terms of Service. With the help of our community, we’ve been able to review and enforce our policy on dozens of sites.

We appreciate all of the help and feedback we’ve received and we are excited to continue to partner productively with users and organizations that are passionate about fighting toxic content and hate speech. To improve our efforts, we’ve built a Terms of Service Violations Submissions form. This form is a way for users to explicitly share with us when they’ve found a community to be in violation of our terms. In addition to reporting individual users (which helps moderators know who in their community is perhaps exhibiting toxic behavior), you can now report directly to us when you think there’s a publisher/site we should take a look at. When we are made aware of potential violations, we review them internally and make a decision about whether or not to allow the site to remain on our platform.

This isn’t a small scale matter; we know that to have a meaningful impact across our network, we need to build solutions into the product. With that in mind, we’re committed to building tools to make the moderation experience easier and better for publishers (and commenters, too).

Here are some things that we’re working on:

  • More powerful moderation features. We’re working on two features right now, Shadow banning and Timeouts, that will give publishers more options for managing their communities. Shadow banning lets moderators ban users discreetly by making a troublesome user’s comments only visible to that user. Timeouts give moderators the ability to warn and temporarily ban a user who is exhibiting toxic behavior.

  • Toxic content detection through machine learning. We are working on a feature to help publishers identify hate speech and other toxic content and then handle this more effectively.

  • Commenting policy recommendations. While we already provide suggestions for how to create community guidelines, we’ve realized that we can be more proactive and more assistive to our publishers. We’re working on helping our publishers expose their custom commenting and community guidelines by making them more visible to their readers and commenters.

  • Advertiser tools: Just like publishers do not want toxic content on their sites, we know that advertisers do not want their content to display next to toxic comments. Leveraging our moderation technology, we will provide more protection for advertisers, giving them more control over where they display their content.

If you think this is a simple problem to solve, you are sadly wrong. And if you think that Disqus and USV and I don’t care about solving this problem, you are wrong about that too.