It only took 36 hours for these students to solve Facebook's fake news problem(businessinsider.com)
businessinsider.com
It only took 36 hours for these students to solve Facebook's fake news problem
http://www.businessinsider.com/students-solve-facebooks-fake-news-problem-in-36-hours-2016-11
7 comments
This article itself felt like fake news being that it states they didn't solve the fake news problem: "A Chrome plug-in that labels fake news obviously isn't the total solution for Facebook to police itself."
> This article itself felt like fake news
Totally impossible, I say.
Just how hard of a problem is it for an algorithm to determine real news from lies?
Not that hard.
... hackathon ... just 36-hours ... classifies every post ... artificial intelligence ... algorithm ... reasonable certainty ... such WOW ;)
And... how exactly does their AI tell apart two election articles differing only in reported results? What about determining whether alleged Linus Torvalds' statement on penguin equality is real or not? Or whether some random hackathon project actually delivers, for that matter?
Why did they write an article like that? What about people who don't know that this project (most likely) reacts only to cranky language or some finite number of facts it has been trained on and could be easily gamed? They'll go and share this crap to "prove" that Facebook is lazy if they feel like doing so, that's what. So thank you, BI, for your contribution to the quality of Internet discourse.
Totally impossible, I say.
Just how hard of a problem is it for an algorithm to determine real news from lies?
Not that hard.
... hackathon ... just 36-hours ... classifies every post ... artificial intelligence ... algorithm ... reasonable certainty ... such WOW ;)
And... how exactly does their AI tell apart two election articles differing only in reported results? What about determining whether alleged Linus Torvalds' statement on penguin equality is real or not? Or whether some random hackathon project actually delivers, for that matter?
Why did they write an article like that? What about people who don't know that this project (most likely) reacts only to cranky language or some finite number of facts it has been trained on and could be easily gamed? They'll go and share this crap to "prove" that Facebook is lazy if they feel like doing so, that's what. So thank you, BI, for your contribution to the quality of Internet discourse.
It's a pretty cool plugin, really. But I'm guessing that the people interested in using it already belong to a demographic that is highly skeptical and good at spotting fake news.
It's a cool hackathon project to play with classification. Presenting it as even a partial solution, as the article did, is irresponsible.
Why not? Facebook doesn't need to filter for you, you can filter however you like. What's wrong with a user controlled solution?
Gee, the real problem here is that none of the people reading the "fake news" consider said fictitiousness as a bug.
Indeed, fake aspect of such so-called-news is the feature for them.
Wrap your head around that.
It's news for the world they wished they lived in.
Indeed, fake aspect of such so-called-news is the feature for them.
Wrap your head around that.
It's news for the world they wished they lived in.
Post fact world. It's all about feelings and baseless opinion now.
Indeed, Newt Gingrich said during the campaign:
> GINGRICH: The current view is that liberals have a whole set of statistics that theoretically may be right, but it's not where human beings are.
> CAMEROTA: But what you're saying is, but hold on Mr. Speaker because you're saying liberals use these numbers, they use this sort of magic math. These are the FBI statistics. They're not a liberal organization. They're a crime-fighting organization.
> GINGRICH: As a political candidate, I'll go with how people feel and I'll let you go with the theoriticians.
Could not be clearer, I suppose.
http://mediamatters.org/video/2016/07/25/john-oliver-theme-r...
> GINGRICH: The current view is that liberals have a whole set of statistics that theoretically may be right, but it's not where human beings are.
> CAMEROTA: But what you're saying is, but hold on Mr. Speaker because you're saying liberals use these numbers, they use this sort of magic math. These are the FBI statistics. They're not a liberal organization. They're a crime-fighting organization.
> GINGRICH: As a political candidate, I'll go with how people feel and I'll let you go with the theoriticians.
Could not be clearer, I suppose.
http://mediamatters.org/video/2016/07/25/john-oliver-theme-r...
The binary classifying could be an issue there, additionally I love the current trend of using artificial intelligence as a black box no one need to explain.
Nevertheless, "Facebook" and "Fake news" seems very trendy this week.
Nevertheless, "Facebook" and "Fake news" seems very trendy this week.
Yeah that trend is great. AI officially has zero meaning at this point.
"But the students show that algorithms can be built to determine within reasonable certainty which news is true and which isn't" I know this is a puff piece and not real news (ugh ironic) but this is so overstated...
So it gives us the opinion of these students about some websites. Just having heuristics that essentially appeal to the authority of source prestige was a big problem of the election and not invented by Facebook although the invisible isolation it allows can exacerbate it a bit.
Great solution for the end user, but the problem from Facebook's end is far more constrained than 'stop showing fake news.' They also have to find a way to do it without affecting sponsored articles and engagement.
Technical solutions to the problem are relatively cheap, but the root of the problem is much deeper, involving social media's tendency to function as a non-fact-checked echo chamber. Facebook is just the most notable platform plagued by the issue.
In the past, Facebook had a partial solution-- human editors who would curate news stories and try to eliminate fake news from trending. I recently met one of these editors while traveling, and we had some long discussions about the issue. Facebook fired the team of human editors and replaced them with an automated solution. That decision was motivated based on a combination of factors related to expenses and liability which is obscured by Facebook's opaque internal hierarchy.
Really addressing the problem will take significant investment by platforms like Facebook to research the problem and implement robust technical solutions. But these platforms have rigid internal hierarchies and internal feedback loops that prevent that from happening. For platforms to invest in solving this problem, continued, strategic pressure needs to come from the outside.
(As it is, there is already pressure from the outside, but Facebook has responded by basically trying to cover their own ass: http://gizmodo.com/facebooks-fight-against-fake-news-was-und...)
The ex-Facebook curator I spoke with attempted to change Facebook's culture from the inside and found it impossible. Her recommendation is that credible journalism organizations from varying ideological backgrounds (everything from CNN, to Fox News to the New York Times) need to demand that Facebook invest in a solution, by refusing to syndicate their content through Facebook's platform unless Facebook does something to prevent credible, fact-checked content from getting shouted over by sensational fake news.
Journalism organizations are already working on this, but readers can help by writing letters to the lead editors of these organizations urging them to put this kind of pressure on Facebook. People with technical skills can help by investing time and energy improving competing news-syndication platforms (Twitter, Reddit, RSS, Steemit, etc).
I've got my own elaborate technical solution to Facebook's fact-void-echo-chamber problem that involves a combination of labels, like these students created, and algorithms that recommend related stories from sources across ideological divides. But unless leading social media platforms are ready to invest in these kinds of solutions, their impact on the problem will be negligible.
Anyway, these are all just my thoughts after listening to many people talk about the issue. If anyone reading this is a part of the team that's investing energy into getting this problem solved, I'd love to help. I'm currently a JS/React dev and also willing to shift my career to marketing, UI/UX, etc. Please reach out to me: http://josiahsprague.com
In the past, Facebook had a partial solution-- human editors who would curate news stories and try to eliminate fake news from trending. I recently met one of these editors while traveling, and we had some long discussions about the issue. Facebook fired the team of human editors and replaced them with an automated solution. That decision was motivated based on a combination of factors related to expenses and liability which is obscured by Facebook's opaque internal hierarchy.
Really addressing the problem will take significant investment by platforms like Facebook to research the problem and implement robust technical solutions. But these platforms have rigid internal hierarchies and internal feedback loops that prevent that from happening. For platforms to invest in solving this problem, continued, strategic pressure needs to come from the outside.
(As it is, there is already pressure from the outside, but Facebook has responded by basically trying to cover their own ass: http://gizmodo.com/facebooks-fight-against-fake-news-was-und...)
The ex-Facebook curator I spoke with attempted to change Facebook's culture from the inside and found it impossible. Her recommendation is that credible journalism organizations from varying ideological backgrounds (everything from CNN, to Fox News to the New York Times) need to demand that Facebook invest in a solution, by refusing to syndicate their content through Facebook's platform unless Facebook does something to prevent credible, fact-checked content from getting shouted over by sensational fake news.
Journalism organizations are already working on this, but readers can help by writing letters to the lead editors of these organizations urging them to put this kind of pressure on Facebook. People with technical skills can help by investing time and energy improving competing news-syndication platforms (Twitter, Reddit, RSS, Steemit, etc).
I've got my own elaborate technical solution to Facebook's fact-void-echo-chamber problem that involves a combination of labels, like these students created, and algorithms that recommend related stories from sources across ideological divides. But unless leading social media platforms are ready to invest in these kinds of solutions, their impact on the problem will be negligible.
Anyway, these are all just my thoughts after listening to many people talk about the issue. If anyone reading this is a part of the team that's investing energy into getting this problem solved, I'd love to help. I'm currently a JS/React dev and also willing to shift my career to marketing, UI/UX, etc. Please reach out to me: http://josiahsprague.com
> robust technical solutions.
Why do you believe technical solutions are possible? This is not a technical problem.
Why do you believe technical solutions are possible? This is not a technical problem.
Technical solutions alone are not enough, but any solution to the problem will require changes to the technology that is creating the problem in the first place.
Ad-supported social media are driven by engagement, and critical thinking skills are unevenly distributed. Those two problems aren't likely going anywhere.