[sriramk]

This actually tripped up a friend of mine a couple of years ago. She left a comment on a photo of someone holding a toy gun saying "You look like <insert-name-of-well-known-terrorist>" followed by a smiley. Within hours, she got a message and a phone call from someone claiming to be working for FB's security who asked her some basic questions on why she left that comment. The whole experience scared her from using FB for a long time.

I thought the whole thing was adhoc and confusing. Anyone who saw the comment could easily see that it was a joke. Also, if it wasn't a joke, why is FB calling her and not someone from law enforcement?

Would love it if someone from FB here on HN could comment.

[nbarlow]

It sounds like your friend got trolled. We don't call users whom we suspect are criminals, and we certainly wouldn't call someone we suspect of being a terrorist. Especially on the grounds of a terrorism joke - believe it or not, we do have a sense of humor.

I work for Facebook's User Operations team and, as the Reuters article says, this specific tool targets the (thankfully) rare cases of adults trying to use the site for the purpose of grooming kids.

We use advanced technical systems to specifically identify grooming situations and strive for a low false positive rate. We have strong internal controls around these tools to prevent misuse or abuse, and stringent guidelines for the way we cooperate with law enforcement.

For whatever it's worth, I have been at Facebook for several years, and I am so amazed every time we're able to help a child avert an absolute worst-case scenario. These cases are rare, but they do happen, and I'm grateful we have the tools necessary to keep the worst of the worst from unfolding in the real world.

[Produce]

Speaking for the internet and most of humanity here - go fuck yourself and get a real job.

Anyone who argues that everyone should be spied upon in order to protect a minority is an idiot who does not understand the concept of freedom. Privacy is a right.

[pilif]

You shouldn't be striving for a low false positive rate, but for a zero false positive rate.

"Think of the children" doesn't rationalize even one wrongly accused person - especially in today's society where an accusation is enough to completely ruin ones life.

[richardw]

I'm very sure a lot of things have to happen before a weird private message thread ends up in an accusation that ruins someone's life.

A false negative could ruin the child's life just as easily, so I suspect most parents might think that's more important. It's not necessary to choose either "privacy uber alles" or "think of the children". As with most things in life, there's likely a good balance somewhere between the extremes.

[flyinRyan]

>A false negative could ruin the child's life just as easily

For example? A child being preyed on by a pedophile? What would stop that from happening in real life? Maybe we should set up cameras every five feet since obviously if we could stop it (no matter how immoral we have to act to do so) we have an obligation to, right?

[richardw]

Yes, a child being preyed upon by a paedophile. Nothing would stop it happening in real life, why is that relevant? Kids can die from drinking too much water, but using your reductio ad absurdum, you'd accuse me of wanting to ban water. Obviously there's a reasonable balance, which you continue to reject.

People have obviously found that online fora are perfect to groom children, because you can approach a child far easier than doing so in a playground. You can also repeatedly try with different children without your likeness being plastered on a wall somewhere. Since it's an obvious target, and checking it can be automated, why not police it in a balanced way? Again, you only see one side of this.

Similarly, in real life if you had a cheap, automated way of highlighting interactions that have been proved to have a higher chance of leading to children being molested while still balancing that with the human rights of the adult, most would find a balance between "do nothing" and "be hysterical". We don't have such an automated mechanism, and I'm sure your cameras-every-five-metres idea is on the hysterical side of balance, so even using it is obviously a straw man argument. That doesn't discount more rational approaches for those who have more than one principle in mind.

[taligent]

Maybe they live in the real world where a zero false positive rate simply doesn't exist.

And be serious. Getting asked a question from Facebook is hardly going to ruin someone's life is it now ?

[pilif]

OP said they aren't calling users but passing information on to law enforcement. Getting a call by Facebook for a bad joke is one thing. Getting the door busted in by police and subsequently shamed and fired from your job because of a bad joke you thought you made in private is another.

in case of law enforcement, the only acceptable false positive rate is zero.

[jiggy2011]

The only way to achieve this would be to have no law enforcement.

[tatsuke95]

>"a zero false positive rate simply doesn't exist."

Laughable. It does when Facebook are going out of their way to read something that is absolutely none of their business. You can easily obtain zero false positives by not turning anything over the the police.

[flyinRyan]

> I am so amazed every time we're able to help a child avert an absolute worst-case scenario. These cases are rare, but they do happen, and I'm grateful we have the tools necessary to keep the worst of the worst from unfolding in the real world.

Example? Because it sounds like justification bullshit.

[_debug_]

OP wrote in another message that "her account got temporarily disabled within a short time of her posting that comment". Is that possible for non-Facebook employees to effect somehow? If not, then it was one of your colleagues, not a troll.

[sriramk]

I'm the OP. It could be possible that someone knew how to get into her account and had just done a self disable/enable. Far fetched though.

[HeyLaughingBoy]

Considering how dumb most people's passwords are, it's likely easier than you think.

[personlurking]

Reminds me of Apple sec. pretending to be police in order to track the missing iPhone.

[antidoh]

Never forget.

[qq66]

It could also be a "friend" who saw the status update and had a sick sense of humor...

[rationalbeats]

How could the friend even find a way of contacting someone at facebook. That is near impossible.

[slig]

I think he meant the friend called her pretending to be from FB, as a in phone prank.

[sriramk]

Nope, her account got temporarily disabled within a short time of her posting that comment, got unlocked after a day or two (I'm not sure how long it took to become available). Sure, some friend could theoretically have found a way to get into her account, done the locking and unlocking but seems pretty far fetched.

[gojomo]

There's almost always been some form of 'flag'/'report as inappropriate' button.

[flyinRyan]

Why didn't your friend just hang up on them? If someone is that stupid, don't give them the time of day. Though, I might ping my lawyer after laughing at them and hanging up just so someone knows if I suddenly disappear.

[einhverfr]

we should have a day of protest doing exactly this to eachother......