Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Facebook admits it must do more to stop the spread of misinformation

techcrunch.com

Post reply on HN
Print view View original

91–100 of 645 posts

Re: Facebook admits it must do more to stop the spread of misinformation

#91

Post by wyager »

denzil_correa wrote: post #79
btilly wrote: post #11

This article criticizes Facebook for firing the human editors that had been keeping things sane. However they were pushed to that by accusations of bias in the right-wing media. Accusations that looked likely to lead to a Congressional investigation. See, for example, http://thehill.com/policy/technology/279361-top-republican-d... . Now they are in a situation where they are damned if they do, damned if they don't. A…

> Now they are in a situation where they are damned if they do, damned if they don't. And people immersed in echo chambers will accuse them of bias no matter what. You are making an assumption that algorithms can not be biased. Earlier, humans were biased. Now, algorithms are biased.

Provided the data isn't wrong, we can typically put tight statistical error bounds on the output of various ML algorithms. Most people's understanding of how algorithms can be "biased" is completely wrong. There was a whole furor a while back about algorithms discriminating due to language differerences in the data for people of different races. Guess what; if you give race as a parameter to any half-decent algorithms trained on that data, the algorithm will learn that the written data contains those biases. Decent algorithms trained on true data are pretty much guaranteed to have results within a very very small margin of reality.

Re: Facebook admits it must do more to stop the spread of misinformation

#92

Post by grzm »

Spellman wrote: post #87
AlexC04 wrote: post #76

I think it would be relatively easy to have an "auto-snopes" feature which detects a URL being shared and immediately attaches a post that says "this has been debunked by X" For example - I've seen dozens of links to Michael Moore's trumpland speech which strategically ends with the phrase "America will elect Trump and it will feel great" Sample: http://www.zerohedge.com/news/2016-10-25/michael-moore-trump... He conc…

I have personally asked an individual why they don't trust snopes. Their claim is that it's a Democrat rag towel and shill. I then ask, then what fact-checker would you trust? Their response: None of them. Congratulations. We've reached a point, for this individual, where "facts" as decided in the common forum are suspect and the only thing they trust is themselves (and whatever non-Mainstream Media they listen to).…

This is something that concerns me a lot as well. How do we rebuild this?

Re: Facebook admits it must do more to stop the spread of misinformation

#93

Post by nostrademons »

jerf wrote: post #27
btilly wrote: post #11

This article criticizes Facebook for firing the human editors that had been keeping things sane. However they were pushed to that by accusations of bias in the right-wing media. Accusations that looked likely to lead to a Congressional investigation. See, for example, http://thehill.com/policy/technology/279361-top-republican-d... . Now they are in a situation where they are damned if they do, damned if they don't. A…

Yes, trying to solve this problem by creating a definition of "misinformation" sufficiently precise to act on it for all possible articles, then trying to remove it somehow is probably an AI-complete problem. If it's not AI-complete it's probably just plain an ill-formed question. That's before we ask questions about the biases that get embedded into the misinformation detector. This can't be fixed at Facebook scales…

This is a problem with human wetware - there are several cognitive biases [1][2][3][4] that make people engage more strongly with ideas that fit their preconceptions and discount ideas that don't. Add information-sharing to the mix and the aggregate effect you get is large information cascades that fracture the population into tribes that believe things not necessarily because they're true, but because they are stated early, vehemently, and frequently.

Remember that evolution favors individuals who survive & gain status, not those whose representation of the world is true. Someone who is right but killed for it (eg. Copernicus) is still dead.

The fix is for people to be aware of these cognitive biases in themselves, to actively seek out information that differs from their preconceptions, and to challenge inaccurate information (with evidence!) when presented with it. You can't outsource this to a communication platform, because it's inherently less comfortable and more effort for the reader, and the reader will just go to a competing communication platform that makes them feel better.

[1] https://en.wikipedia.org/wiki/Illusory_truth_effect

[2] https://en.wikipedia.org/wiki/Confirmation_bias

[3] https://en.wikipedia.org/wiki/Bandwagon_effect

[4] https://en.wikipedia.org/wiki/Congruence_bias

Re: Facebook admits it must do more to stop the spread of misinformation

#94

Post by Florin_Andrei »

jerf wrote: post #27
btilly wrote: post #11

This article criticizes Facebook for firing the human editors that had been keeping things sane. However they were pushed to that by accusations of bias in the right-wing media. Accusations that looked likely to lead to a Congressional investigation. See, for example, http://thehill.com/policy/technology/279361-top-republican-d... . Now they are in a situation where they are damned if they do, damned if they don't. A…

Yes, trying to solve this problem by creating a definition of "misinformation" sufficiently precise to act on it for all possible articles, then trying to remove it somehow is probably an AI-complete problem. If it's not AI-complete it's probably just plain an ill-formed question. That's before we ask questions about the biases that get embedded into the misinformation detector. This can't be fixed at Facebook scales…

It's almost as if solving the problem of truth is antithetical to the notion of having a for-profit entity be the main channel of [legit|mis-]information in our world.

It's almost as if the way we understand and use money has hardened into a giant screwball of lies, base instincts, and perverse incentives.

Almost.

Re: Facebook admits it must do more to stop the spread of misinformation

#95

Post by AlexC04 »

youdounderstand wrote: post #12

Just stop using aggregators. Pay for real news sources that hire actual investigative journalists.

This is actually an incredible option ... and for me, quite a mind-blowing one.

Could news.facebook.com be an actual content-producer with real journalism?

Large, investigative projects on what is possibly the world's largest distribution platform.

Re: Facebook admits it must do more to stop the spread of misinformation

#96

Post by Florin_Andrei »

nostrademons wrote: post #93
jerf wrote: post #27

Earlier quoted context omitted.

Yes, trying to solve this problem by creating a definition of "misinformation" sufficiently precise to act on it for all possible articles, then trying to remove it somehow is probably an AI-complete problem. If it's not AI-complete it's probably just plain an ill-formed question. That's before we ask questions about the biases that get embedded into the misinformation detector. This can't be fixed at Facebook scales…

This is a problem with human wetware - there are several cognitive biases [1][2][3][4] that make people engage more strongly with ideas that fit their preconceptions and discount ideas that don't. Add information-sharing to the mix and the aggregate effect you get is large information cascades that fracture the population into tribes that believe things not necessarily because they're true, but because they are state…

> The fix is for people to be aware of these cognitive biases in themselves, to actively seek out information that differs from their preconceptions, and to challenge inaccurate information (with evidence!) when presented with it.

You are right, but I hope you are aware that this would require a giant leap to an essentially superhuman state of rationality and clear thinking for a vast majority of people.

Just stop for a moment and look around you. Is that how most people actually function in this world?

Re: Facebook admits it must do more to stop the spread of misinformation

#98

Post by debt »

AlexC04 wrote: post #76

I think it would be relatively easy to have an "auto-snopes" feature which detects a URL being shared and immediately attaches a post that says "this has been debunked by X" For example - I've seen dozens of links to Michael Moore's trumpland speech which strategically ends with the phrase "America will elect Trump and it will feel great" Sample: http://www.zerohedge.com/news/2016-10-25/michael-moore-trump... He conc…

the problem is the deeper fundamental philosophical conversations about truth and the meaning of truth and relativism and biases etc aren't being discussed at in the tech community let alone at FB/Twitter HQ.

Having a lil fact checker isn't as easy as adding a button to snopes.

There's bias in language itself and how even a sentence is structured. But again a deeply complex conversation nobody seems to be having in this community.

Re: Facebook admits it must do more to stop the spread of misinformation

#99

Post by faet »

trynumber9 wrote: post #20

You think the users are too stupid to fact check. But what makes you think the users care about the truth anyway? More often than not they double down with "well it's emblematic of the greater problem" or something to that effect. They then look for new evidence to support their views. Like it or not, some sites are designed to build echo chambers and are not for general discourse.

A lot of people don't care. They believe an image on facebook that reaffirms their belief.

You try to counter with statistics from the FBI, DOJ, BLS, DOL, or any other organization and the numbers are 'made up'. And studies have been done that when presented with counter arguments it strengthens the original misperceptions.

Corrections also don't get much traction. I've seen many conspiracy theories pop up for a week, then die down. While the correct gets no traction at all.

>As a result, the corrections fail to reduce misperceptions for the most committed participants. Even worse, they actually strengthen misperceptions among ideological subgroups in several cases.

http://www.dartmouth.edu/~nyhan/nyhan-reifler.pdf

People don't want news, they want to hear what they already 'know'.

Re: Facebook admits it must do more to stop the spread of misinformation

#100

Post by whybroke »

btilly wrote: post #11

This article criticizes Facebook for firing the human editors that had been keeping things sane. However they were pushed to that by accusations of bias in the right-wing media. Accusations that looked likely to lead to a Congressional investigation. See, for example, http://thehill.com/policy/technology/279361-top-republican-d... . Now they are in a situation where they are damned if they do, damned if they don't. A…

> damned if they do, damned if they don't

The level of damnation differs significantly.

Having a human team being accused of bias (possible falsely) and pressured by politicians to change puts them in the same category as the NYT, the Wall Street Journal and every serious print and broadcast service on earth.

Algorithmically spreading exciting and blatantly false claims puts them in the category of the National Enquirer, and 'The Protocols of the Elders of Zion'

One category is significantly less damned than the other.

Post reply on HN