If you were asked to solve the 'fake news' problem, how would you solve it?

[deepakravindran]

I am working on a new idea Shrimp ( domain shri.mp), and our mission is to fix the fake news problem.

We have nailed down the user experience in three directions. I would love your thoughts on these three ideas - 1)Build a secure browser so that it can work across the web and mobile devices which will automatically sync up with an online database of all the fake news articles, so whenever someone clicks on a link, he will be alerted. We initially built a Chrome extension and decided to scale it up using Firefox source code.

2)Develop a software like an antivirus which scans every link that you receive through social media especially via messengers such as WhatsApp and Fb where they spread the fake news. There is also a synergy that I see using existing networks such as Reddit or Jelly user's to vote down the new articles which are fake and not just sending out patches or updates.

3)Build a rating system for journalists where we can rate the authenticity of articles based on that like NY times will embed our code just like how they do for like button - it will act like a trust score.

Do let me know your thoughts :)

[philiphodgen]

"How do you 'solve' fake news?" is a question fraught with peril.

What does "solve" mean?

What does "fake news" mean?

Is it possible that opinions may vary on what these two things mean?

On these points, reasonable people of good will can have differing points of view.

In my more cynical moments, however, I hear the echoes of the beginnings of the war on drugs. Except far more frightening -- because this is a war on thinking. By people who sincerely think they are doing a noble act.

If you want to do something about the problem, cut demand. Make smarter readers.

[pcvarmint]

You can't solve it anymore than you can have 100% "objective" news.

People need to be free to read all viewpoints and come up with their own opinions.

Most of the time, news is called "fake" because it says things which, in the context of our preconceived values, leads to contradictions or absurdities. Viewed from a different value system, the same facts can have totally different meanings and consequences. Other times, "fake news" is just smear tactics for a larger political agenda.

A site claiming to identify or prevent "fake news" just winds up being another opinion site.

[NeutronBoy]

I agree. You need to be able to differentiate between 'types' of fake news. It means different things to different people. Three perspectives I can think of at the top of my head, but it's a sliding scale:

1. Flatout lies with no factual basis (Hillary Clinton/Donald Trump is a lizard person from Mars)

2. Mis-truths of accepted facts, or mis-representation of statistics (climate change)

3. Opinions that you don't agree with (ACA is good/bad, or The Media is good/bad)

[sheepmullet]

We are already very good at detecting and routing around #1.

The problem is people use #2 and #3 to build up ideologies in the same way that advertisers build up brands.

Take for example the wage gap between genders.

There is a moderate wage gap (20-30%) if you look at unadjusted figures.

Fringe feminists try and use this stat to build a general sense of unease with their followers.

This primes the reader for a few anecdotes of women being unfairly passed over for promotion and the reader walks away thinking there is a widespread issue of women being treated poorly in the workplace.

White nationalists do the same thing. It is a fact that African Americans make up a disproportionate amount of murderers by race.

White nationalists will then use a few anecdotes of African Americans killing innocent white families and the reader walks away with a negative impression of African Americans.

And now see what I've just done with my comment: I've linked white nationalists and feminists.

By placing these two examples together I am connecting them for the reader and subtly discrediting these feminists.

[20years]

My biggest concern is who determines what is fake news? Seems as though people are now shouting "fake news" if a story doesn't fit within their belief system or agenda.

I don't think a voting platform is the solution either because you then get the herd mentality and it can so easily be gamed.

I am not even convinced that "fake news" is the issue. I am more on the side that our society has lost its ability to truly evaluate what they read or have any real level of critical thinking at all. If more of our society questioned what they read or even took the time to truly read things, I am not so sure this would be an issue. Rather, so many simply go by what an excerpt says in their Facebook or Twitter feed or what their followers are saying.

[pizza]

Cryptography - centralized/decentralized certification of special <article news-UUID=<some long hash> > elements. User can install cert authorities like [NYT, RT, Reuters] or whatever their preference, who serve as independent consensus auditors of sorts. To prevent leakage of which user is reading which article, maybe repurpose some kind of provably-anonymous voting scheme like [0]. Then use the look up of the hash of the article to name it

[0] https://arxiv.org/pdf/1702.05000.pdf

[jelliclesfarm]

I am not a technical person. This is a non coder/lay person's point of view, okay?

With that disclaimer, there are some things you have to do: 1. Determine a lie detection strategy. Memes mutate. Tracking the way news changes and differs ever so slightly would require some kind of crawler bot that constantly compares and contrasts several versions and iterations of the same news. Can real news mutate? Probably not. Fake news would keep changing with each pattern because that's it's purpose..to spread and it would use every advantage it can find..2. Pattern recognition. There is a certain pattern to how fake news channels spreads before it comes main stream. 3. Reputation system or rankings. 4. Humans develop deception detection methods as we get more exposed to more deception. So to detect fake news, expose your (what I can only imagine) code or bots or whatever to fake news. 5. Karl popper's falsifiability. Science has the demarcation problem to distinguish between science and pseudoscience. I would like to think of it as science vs unexplained phenomena or non-science.(example: I don't think homeopathy is pseudo science and am reluctant to label it as such..I am also reluctant to label it science) Study scientific method in theory before getting into technical code.

[stephancoral]

The most effective and proven way to deal with "fake news" is to establish a Ministry of Truth.

[deepakravindran]

"Ministry of Truth." Love it! :)

[RikNieu]

I'm afraid that you might need to validate this idea before jumping in and spending a lot of time and money on it.

The problem is people who fall for and spread fake news don't want to be told it's fake news, they just want to have their world views validated.

[deepakravindran]

Agree with your point on how people fall for it - but the most important thing we are trying to solve here are the actual credible facts - and not opinions of those facts. :)

[rchaud]

After attending a recent hackathon focused on this topic, I came away thinking most of the solutions were emphasizing some kind of automated fact checking solution.

I don't know if that alone will fix it. Fake news exists because people want to believe it, and are willing to lower their bullshit meters in order to use the article as "ammunition" in a political argument with their friends/family/Twitter egg. People aren't looking for "antivirus" tools to stop fake news; they're actively looking to share links that confirm their view of the world. It doesn't even matter if the link has a "wordpress.org" or "blogspot.com" domain. We live in a time when people don't even read the articles, they just look at the headline and jump into the comments to fight someone.

To fight fake news effectively, perhaps we need to do a better job of uncovering why people apparently have so much time to spend on the Internet to argue with people they'll never meet. Perhaps while we're at it, we could also look into why we have a political system where legalized bribery ('lobbying') gives preferential access to the powerful. We're at the point where winning elections is based less on getting enough people to agree with your vision, and more on playing one group of disaffected voters off against another.

I wish you luck, but I believe this is a people problem, not a tech problem.

[wazanator]

Rate the writer not the publisher. Get your users interested in the writings of individuals instead of who they write for. News providing companies are like any other, some workers are lazy and others actually really care.

If Bob is posting fake news block him instead of blocking company XYZ because Cindy who also works for XYZ actually puts out good articles.

Be careful of straying into defamation territory. Don't name and shame. For lack of a better term shadow ban the writer.

[grzm]

I agree that keeping track of writers is important. I currently don't have an easy way to do that. Anyone aware of a source that rates journalists in a centralized manner?

I think it is important to keep in mind the publisher, as they do have a responsibility for choosing the articles and journalists they publish. If it becomes clear that a publisher has a tendency to publish low quality (for various measures of quality), I think it does make sense to take that into account for anything else the publisher may publish.

[wazanator]

Also I would look at how Wikipedia handles news. Not many people are probably aware of it but they do post current news on the front of the site. It's slower to come out then traditional news but it is a good source because of how well they go about citing sources.

[dimodi9]

I believe one of the hardest elements here is to battle with the incentives of those that publish such news. In effect a "crowd" based approach could help to distinguish signal from noise, that said, fake-new publishers will inevitably use bot-nets to counter act any crowd-based check.

Aside from human curation by a trusted group - which is extremely expensive - the only solution I can think of is a share-profit system to be implemented in already trusted news services and brands, so that more people can monetize their news articles. That way, you can use existing journalists to validate all news, and their brand to distribute, while providing incentives to niche-authors.

Any technological solution like the ones you are describing 1-3 can be attacked.

[deepakravindran]

Thanks. Yes, the incentives should be based on building yourself a credible profile on the internet - like a trust button. A lot of community-based websites including Reddit, Jelly and others are moving in this direction but the challenge is how fast a community can move and how they can stop it in other networks such as facebook and whatsapp where this spread.

[ev_rolfe]

I don't think this is something which can be solved by some "killer algorithm".

It seems that this whole fake-news thing is largely because of hacker news-ey type people wanting to use algorithms to make everything better and introducing them into as many aspects of our lives as they can.

Maybe then the solution isn't more algorithms to fix what the existing algorithms are doing wrong but changing the culture of news sources to encourage more stringent journalistic standards? This would require a lot of hard work amongst many people in influential position but most problems are not solved by coming up with the right code.

[f_allwein]

I think a lot of this could simply be done on a domain level, i.e a site that publishes fake news once is likely to be generally less trustworthy. Then it would be up to sites like Google and FB to remove them from their listings (as is already done with spam) and not accept any ads linking to them (as is done for all kinds of illegal goods).

Come to think of it, even a site's PageRank should be a useful signal, as it (ideally) reflects the trust other sites have in this site.

[deepakravindran]

Our initial try was that by building a Chrome plugin called 'fakeornot' which will just tell you if its a fake site or not. But this was not working accurately as we saw sites like Mashable and others write articles which are not even verified and they quote to a source.

[babyrainbow]

Use a very limited version of the english language. Something like the one used in sending Telegrams in old days. Also, use same size fonts for all the news.

Also, put the details of the news behind a NLP software and let the users ask it more details about a certain news and let it answer then in one or two words...

[nnn1234]

Applaud your effort. The Comment from dimodi9 does make sense. I would think long and hard about the incentive structure.

What do you think of initiatives like Steem?

The Problem of using the crowd to annotate the news is a secondary solution is good but having a source of truth is the final solution

[deepakravindran]

The route of community building is interesting and I can see many people following that. If you see the latest development at Biz Stone's Jelly, their vision is to make a trust score on users and build credibility. The challenge is how quick a community can respond. Fake news is growing like the snowball effect the more it spread on the social media the more it becomes a problem. We need to put a filter before we share the news inside our network - like a verified posting :)

[nnn1234]

One little push back, someone had to write the fake news for other people to share it.

Massive click farms and hate baiting are the issues that can be solved. Options like where the content originated from, who the writer was and why are they anon if so?

Tech solutions are a first approximation to solve this question. Deeper solutions require scale and some mustache twirling

[DrNuke]

Fake news = never happened is different from fake news = exotic (stretched, unconsequential, illogic, etc.) point of views. The former may be technically approachable, the latter I fear we technically can't.

[jenkstom]

Isn't this what the "Web of Trust" issue addresses? Sure, it's complex and would take a lot of time to understand and propagate, but it's a possible solution.

[itamarst]

A lot of sites accused of being fake news are not. Be prepared to be sued for defamation, quite possibly with good reason.

[KanyeBest]

>our mission is to fix the fake news problem.

Sincere question: Why is it a problem?

[deepakravindran]

the whole internet and social media in general is getting polluted because of this.

[BentFranklin]

Perhaps some sort of blockchain technology

[erik998]

Pay more taxes and strengthen public schools. Push public school education as John Dewey intended, for the betterment of civic society.

https://en.wikipedia.org/wiki/John_Dewey#On_education_and_te... """ According to Dewey, the emphasis is placed on producing these attributes in children for use in their contemporary life because it is “impossible to foretell definitely just what civilization will be twenty years from now” (Dewey, MPC, 2010, p. 25). However, although Dewey is steadfast in his beliefs that education serves an immediate purpose (Dewey, DRT, 2010; Dewey, MPC, 2010; Dewey, TTP, 2010), he is not ignorant of the impact imparting these qualities of intelligence, skill and character on young children in their present life will have on the future society. While addressing the state of educative and economic affairs during a 1935 radio broadcast, Dewey linked the ensuing economic depression to a “lack of sufficient production of intelligence, skill and character” (Dewey, TAP, 2010, p. 242) of the nation’s workforce. As Dewey notes, there is a lack of these goods in the present society and teachers have a responsibility to create them in their students, who, we can assume, will grow into the adults who will ultimately go on to participate in whatever industrial or economical civilization awaits them. According to Dewey, the profession of the classroom teacher is to produce the intelligence, skill and character within each student so that the democratic community is composed of citizens who can think, do and act intelligently and morally. """

We have many people that were taught to trust the mainstream news organizations. Twenty years ago no one would have thought of blogs, news summary sites, gossip sites. We have a large proportion of society that never used library index cards, book glossaries, or used more than one source to study anything. They take any news given as true because it derives from some source of authority.

A child's education should not be focused on making money in the future. It's to create individuals that can function in a democratic society.

Yes you can have links and rate journalists but that won't solve the problem. I can legally change my name to Tom Brokaw and then what would you do... The issue has a striking correspondence to the difficulty of using the OpenPGP Public Key Infrastructure... Sure you can sign keys of journalists and have them sign their articles but it does not mean what they wrote is true. What if their editor took their key/password and published in their name? Or if upon expiration of key they just make another key using some journalist's name? You really want readers checking key fingerprints?

Maybe you can present the various sources of information and try to show links to research notes and such but I never see that provided on any news sources. Oh and what about anonymous insider tips... I see that often in the NYTimes... Why would your source of Truth in the times "Trump" my source of truth...?

Jay Rosen http://pressthink.org/ has struggled with this as well. I think he is for journalists just doing their jobs. He had an interesting discussion on the Bill Moyers show you should watch.

http://pressthink.org/board/

Journalists can only do so much. People need to take responsibility as well. In the 1990s would you take someone seriously that said "Hillary Clinton adopts an extraterrestrial"? http://www.motherjones.com/media/2015/04/weekly-world-news-c... The articles might not be so outlandish now but give enough doubt to the mainstream coverage to confuse most people.