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

Ask HN: Anyone interested in building tools for showing bias in news?

news.ycombinator.com

Post reply on HN
Print view View original

41–50 of 163 posts

Re: Ask HN: Anyone interested in building tools for showing bias in news?

#41

Post by return0 »

harigov wrote: post #34
return0 wrote: post #29

Earlier quoted context omitted.

Hmm, i think its a very difficult task. I think perhaps a useful thing would be to intentionally break the bubbles. Something like an inverse-recommendation engine. E.g. instead of "people who read article X also read Y", you say "people who read X did not read Z". Or if you had access to people's facebook likes (does facebook provide an API?) you could find the most anti-correlated users and recommend each other's l…

Isn't what I mentioned sort of similar? It is difficult for sure which is why I am asking for help.

I think people in general avoid the other side, and they don't have enough time for that. Thats why i suggest a mix up of sides. That way you read each story once, but as a whole you get exposed to both contexts.

Re: Ask HN: Anyone interested in building tools for showing bias in news?

#42

Post by rkayg »

Great idea! I am working on pushing my Facebook colleagues to work on this as well. I think the problem will not be solved unless Facebook makes a serious effort. The issue is that like the filter bubbles that are created in the social network, the company is also in an echo chamber of its own.

So, I think for all the folks that know Facebook employees, break them out of their bubble (if they are already broken, commend them and encourage them to improve Facebook). Show them that Facebook is not so rosy colored as it claims, and that they have a responsibility to build tools that promote truth and inclusion.

Re: Ask HN: Anyone interested in building tools for showing bias in news?

#43

Post by harigov »

michaelbuckbee wrote: post #37

I'm concerned that your frame for this might be off as I don't think the "filter bubble" is rational opposing viewpoints around a particular topic so much as there just being so much misinformation and plain weirdness out there. There's a saying about it being 10x harder to refute bullshit than it is to spew it. How do we in the information and technology wing of society build tools to deal with that? Alex Jones had…

I agree. It looks like this problem is multi-faceted. Maybe the problem that you mentioned is 60% prevalent and what I mentioned maybe 30% (random numbers). We should still try to solve these problems. What do you propose? I am genuinely interested because this is getting out of control.

Re: Ask HN: Anyone interested in building tools for showing bias in news?

#44

Post by mtberatwork »

michaelbuckbee wrote: post #37

I'm concerned that your frame for this might be off as I don't think the "filter bubble" is rational opposing viewpoints around a particular topic so much as there just being so much misinformation and plain weirdness out there. There's a saying about it being 10x harder to refute bullshit than it is to spew it. How do we in the information and technology wing of society build tools to deal with that? Alex Jones had…

Indeed, if we do not agree to a common set of facts and a shared reality, how then do we proceed with any kind of meaningful debate? To paint a basic picture, how do you debate the feasibility and merits of sending humans to Mars with a person who denies the moon landings? You might say these are extreme cases, but with the rise and "normalization" of Alex Jones-type media outlets, this will only get worse.

Re: Ask HN: Anyone interested in building tools for showing bias in news?

#46

Post by harigov »

rkayg wrote: post #42

Great idea! I am working on pushing my Facebook colleagues to work on this as well. I think the problem will not be solved unless Facebook makes a serious effort. The issue is that like the filter bubbles that are created in the social network, the company is also in an echo chamber of its own. So, I think for all the folks that know Facebook employees, break them out of their bubble (if they are already broken, comm…

I think it would gain much more credibility if this were done in the open in a transparent manner. If there is a perfect fit for open source software, this is it. If folks from facebook would like to invest time/effort that would be great!

Re: Ask HN: Anyone interested in building tools for showing bias in news?

#47

Post by PaulHoule »

Another issue is that many biases are structural, technological or driven by commercial pressures.

For instance, I would say CNN is biased toward coverage of school shootings and airplane crashes. CNN has the problem that there is not enough news to fill 24 hours so they run a heavy rotation of the same crap that is cheap to produce. Probably the best footage they show is stuff they downloaded off Youtube.

When you catch the CNN crew on a slow news Sunday they will admit that their problem is engaging an audience, both in the sense that they need to make money and also in the sense that they have some duty to inform the populace, the populace has duty to inform itself, etc. The truth is their content is boring, depressing, and awful but they have varied their formula a lot and they really believe they've found a local maximum of what people will watch.

In some sense CNN was biased towards Trump because he's interesting. I would look for news about Trump every day because it was likely he would say something crazy again and I think this was the case for a lot of other people. CNN, Fox News and MSNBC all had great ratings this season.

This 1971 book

https://www.amazon.com/Information-Machines-Ben-H-Bagdikian/...

is about as ahead of it's time as Ted Nelson's work and is very much about what news would be like in the age of the World Wide Web and it contains a damning indictment of the very concept of "news". (i.e. not only is there not enough news to fill a 24 hour tv show, but it's arguable that there is enough news to fill a newspaper every day)

Re: Ask HN: Anyone interested in building tools for showing bias in news?

#48

Post by crimsonalucard »

How do separate bias from truth?

If I said more asian people are smarter than white people I am biased and racist.

If I said more asians have black hair than white people I am not biased, I'm stating an objective fact.

The only fundamental difference between the two statements is that there are hard numbers lending support to one statement (asians having black hair) and the other statement does not. Neither statement, from a technical standpoint, can be verified definitively.

To build a machine that identifies whether or not a statement is biased one must first build a machine that identifies whether or not the underlying statement is true or false.

Building such a machine is an impossible endeavor because the means in which we identify whether or not something true or false is through data, a source which in itself can be biased.

Re: Ask HN: Anyone interested in building tools for showing bias in news?

#49

Post by sfrailsdev »

I think it would be best to start with a small feature. For example, build a classification tool seeing if perspectives of both sides of an issue are interviewed and quoted. Open source it along with your dataset.

We need to build more composable machine learning based tools, and then we can use machine learning on their results to determine which serve which purposes best.

Re: Ask HN: Anyone interested in building tools for showing bias in news?

#50

Post by harigov »

return0 wrote: post #41
harigov wrote: post #34

Earlier quoted context omitted.

Isn't what I mentioned sort of similar? It is difficult for sure which is why I am asking for help.

I think people in general avoid the other side, and they don't have enough time for that. Thats why i suggest a mix up of sides. That way you read each story once, but as a whole you get exposed to both contexts.

Interesting. I'll keep this in mind.
Post reply on HN