Facebook Stops You From Posting ‘Irrelevant Or Inappropriate’ Comments

This is an isolated incident. "Irrelevant or Inappropriate" sounds like a euphemism for spam- if Facebook is fighting comment spam then I applaud them. Their algorithm just isn't quite right yet. No story here.<p>"Is this censorship?" Are false positives on a spam filter censorship?

Google+ needs to do this. I follow Linus Torvals and some other big names in tech, and there are dozens of worthless comments on every single one of his posts.<p>"+1"<p>"Wow,,,,,,,,,,,"<p>"。。。。。。。。。。。。。。。"<p>Incredibly low signal-to-noise, which makes me not even want to look at the comments.

This is a great example of machine learning gone wrong. Facebook turns the knob on their "positivity" algorithm a little too far to the right.<p>Machine learning is one of this things that works great until it doesn't, and you get misguided decisions like this that had zero human involvement.

This is consistent with Facebook's established strategy. Ultimately Facebook want to make sure that you (the user) will always associate Facebook with positive emotions, not negative ones.<p>It's the same reason that there is no dislike button, in spite of years of requests to the contrary.<p>If you really look closely you will find that Facebook does not have the most advanced UI, nor does it offer the most functionality. People don't use FB for any of those reasons, even if they think they do. People use FB because FB makes them feel happy, connected, and valued among their "friends".<p>Allowing FB to be used to post negative comments runs contrary to these objectives and, due to the current limitations of machine interpretation of language, it is inevitable that some innocuous comments get misinterpreted along these lines.<p>I'm not saying that I am supportive of this practice, but I do understand why they are doing it.

Based on the anecdotal comments below the original post, it seems like there's a very high correlation between posts that get flagged as inappropriate by Facebook and those that mention Google+ or link to it. Which seems absurd, on the face of it, but I am going to have to experiment.

The solution is pretty straight forward. Stop using FB to handle your site's comments. I don't even see comments on TechCrunch anymore because I have FB whitelist only and I'm not the only one.

@nikcub I think it's not a coincidence.<p>In fact FB declines Post and Messages for a while already. For example you can't post or comment and include a picture from imgur, while it (obviously) fine from other hosts. Thats what i experienced so far, but i can imagine there are other examples.<p>That does not need to be censorship. Might be that they got an agreement about not hotlinking or hosting imgur images for copyright reasons.

Facebook's algorithms seem on their way to becoming the 'Allied Mastercomputer'/'Adaptive Manipulator' that tortures the remnants of humanity in Harlan Ellison's dystopic short story, 'I Have No Mouth and I Must Scream':<p><a href="http://en.wikipedia.org/wiki/I_Have_No_Mouth,_and_I_Must_Scream" rel="nofollow">http://en.wikipedia.org/wiki/I_Have_No_Mouth,_and_I_Must_Scr...</a>

This is not isolated. fb's algo sucks. I was posting a genuine comment &#38; tried to post it 3 times. It warned me with a popup, as in the article &#38; the 3rd time, it warned me that i would not be able to comment on public posts.<p>IMHO, a more practical flow would be: warn the user 3 times, but on the 3rd &#38; final time state that : this will be submitted for review by a human reviewer &#38; if found abusive, your a/c will be disable for x days.<p>A genuine commenter would gladly agree to such a request.<p>FB has humans who could review this, or they could use Amazon's Mturk - ship off each abusive comment as a HIT which gets posted after the 3 warning attempts, &#38; maybe send it to multiple reviewers(mturk workers) - &#38; use a majority vote, to decide among the reviews.<p>Decision could be binary ( ban/no ban) or some reputation metric ( decrement by 10 points , &#38; if no suspect activity occurs in the next x(30?) days, karma/ reputation to be restored to the original number.

This is not isolated.<p>I posted a comment on a friend-of-friend's feed. He posted some Bertrand Russell quote and I took issue with the phraseology but generally agreed with the sentiment.<p>When I clicked the "post" button, Facebook gave me some warning message along the lines of "are you sure you want to be posting this? If you post bad things you could be banned from commenting"<p>I found this fairly insulting. I think Facebook has finally found the thing that's going to drive me completely away from their system.

Having just finished a general AI course, this was fascinating and a great example of why to be careful with automation.

The ironic thing is that if Facebook also forbid irrelevant <i>posts</i>, it would become a vast, empty wasteland.

This happens to me fairly often. Though I seem to get this warning more:<p><a href="https://fbcdn-sphotos-a.akamaihd.net/hphotos-ak-snc7/374152_10151486189050161_558110160_23577940_1717231185_n.jpg" rel="nofollow">https://fbcdn-sphotos-a.akamaihd.net/hphotos-ak-snc7/374152_...</a>

So Facebook has decided to take control and accept responsibility for content on third party sites? That's an...interesting choice.

Is it using sentiment analysis?? If so that is very cool. But also annoying; I'm glad I don't use FB.

Completely ridiculous. The next step is Facebook hellbanning people for "inappropriate comments".

People are still using Facebook?

Hey, where’s my “Generate an appropriate comment” button?<p>According to Mark Zuckerberg, people using Facebook are “dumb fucks,” so probably they should be guided?

I'm just super impressed Facebook admins can string two sentences together in a coherent way! I'm happy to give them a piece of gum as a reward. Those fun guys just need to promise not to walk while they chew it.