Twitter to Filter, Warn Users Before They Post

John Lister's picture

Twitter is testing a program to warn users if they are about to post a potentially-offensive message. It's billed as a way to tackle hot-headedness, but could also tackle the limitations of automatic filtering.

At the moment Twitter uses a combination of automated filters that block some content from getting posted, as well as user reports about posts already online.

The test will run for several weeks and cover English language tweets only. It will mean users who are about to post something that involves "offensive or hurtful language" see a message giving them the option of revising it before it goes live.

No 'Banned Words' List Involved

The test won't involve a fixed list of offensive terms. Instead, messages will be compared to Twitter's database of posts that users have previously reported as breaching its policies on acceptable content.

Not all types of post will be included in the initial test; it will only cover people who reply to a previous post. Those are more likely to include targeted, personal abuse compared with more general standalone posts.

Twitter's Sunita Saligram says the company isn't worried about people searching for loopholes in the test to find ways to post offensive material without getting caught out by filters. (Source:

Heat Of The Moment

She says the test is targeted at well-meaning users who simply get worked up when replying to a particular message and don't habitually seek to break content rules. (Source:

As well as making people think again, the test's set-up could reduce the problem of automatic filtering that struggles to deal with context. It's possible some or even many messages flagged for review could actually be fine under Twitter's content guidelines when read in context.

That could give users the chance to either click to confirm the post because they think it's been flagged in error, or to rewrite the post to remove any ambiguity or potentially misinterpreted tone.

What's Your Opinion?

Is this a smart move? Would you be frustrated if you were repeatedly asked to review content you wanted to post? Should sites like Twitter put more effort into outright blocking the worst content, or should they let everything go and leave it to users to report any breaches?

Rate this article: 
Average: 5 (5 votes)


davolente_10330's picture

I have to say that I do NOT indulge in any social media whatsoever and I would be happy with filters that turn the whole lot off completely! The "services" are the invention of the devil incarnate and, from clips I've come across on other sites, seem to turn otherwise mild-mannered folk into hideous, foul-mouthed creatures with the social graces of a bad-tempered grizzly bear. Trump is one glaring example, flinging insults around and hiding behind his Twitter account, which I imagine would not happen face-to-face. HIS account is one that should have been banned long ago, in my opinion, as it serves no useful purpose, other than to show just what a two-faced bigot that man is and is setting an incredibly bad example, that other people are bound to copy, thinking it is the norm.

kitekrazy's picture

Unfortunately this is what happens when Dennis posts this stuff. For those who are not under a rock, there are plenty of Twitter posts that are far offensive. (see celebrities)

If you did not indulge in social media would your TDR show? You'd be surprised that people have used Twitter to defame celebrities, athletes (ESPN), ect. Other politicians have abused this (see Omar) to threaten Israel.

Most of these filters have one intention and that is to silence conservative viewpoints. These are private companies and have the right to filter out whatever they want.

But for some of the reasons you mentioned "social" media can't die soon enough. It's just one big Jerry Springer Show.