Fake news is fake data, 'which makes it our problem', info-slurpers told Data-hungry organisations have been advised to get a better grip on the data they control and work on building trust. In a week when analytic technologies have had more press than ever, many of the discussions at the Gartner Data and Analytics Summit, which ran in London from 19-22 March, focused on dealing with the data …
The main problem is that everyone is talking about "fake news" as if the definition is all out clear, but truth of the matter is that context and perspective are equally important to determine if something is fake, bended or maybe true yet misunderstood.
And then we also have things such as half-truths, politicians love those; simply don't tell the full story so that you're not lying but... in the very sense of the word also not very honest either.
Seriously though, this shouldn't be the problem of the carrier. Nor should those outlets try to make it theirs because that can only lead to one simple outcome: censorship.
Who cares if news is fake or not? If people would stop being so gullible and taking the easy way out then it would be a lot harder to sell this fake news nonsense. Because that's what it is. And even if your news organization is mostly reliable then that by itself is no guarantee that everything you share will be fully true or on par with general perception. I mean... I still remember El Reg carrying that story about how some vague research institute had determined that Microsoft Explorer users had a general lower IQ than Firefox users. Of course within 2 weeks the story got debunked as nonsense, but in the mean time several outlets had carried it and even more funny: several people heavily defended it too.
But yeah, as long as people don't bother themselves to dig a little deeper into what's real and what's "tainted" then nothing really changes. And that's not even touching the issue of opinionated articles.
"The main problem is that everyone is talking about "fake news" as if the definition is all out clear"
The definition is clear.
Those who would like to confuse outright lies with differences of opinion have appropriated it as a catch all term, when it's clearly not.
If a story has no factual basis, it's fake news. This is a world of difference from where certain events are agreed and the causes and consequences are argued over.
So "David Cameron fucked a pig" is fake news, as it is an unsubstantiated rumour, that even those who started it backed away from. Now the media did use appropriate weasel words like "a book alleges DC loves to pork", but it's a good example of something that has been accepted into the popular consensus whilst at no point being true.
Same goes for various reports made in US publications about the perilous state of Europe, where various cities (London, Paris, Amsterdam, Cologne) have been portrayed as being in all out race war with the police unable to maintain public order.
In both these cases, the story is not true, but the target audience really REALLY wants to believe that their views are in fact backed up in reality. So a confirmation that conservatives are feelthy sex pests and that immigrants cause the breakdown of society are what the readers are seeking, and they don't really care if it's true or not.
No idea what to do about it. I'm a very cynical person, in general anything published has an agenda and objective, thus it's always a case of either seeing what they want you to see and thus what to ignore, or reading between the lines when they want to say something, but are being constrained.
Real life is complicated, messy and almost always has really terrible narrative.
If a news story is presented as having a clear narrative that happens to support a particular position (political, philosophical etc), then it usually doesn't pass the sniff test. Same for shitty social research using badly thought through surveys designed to promote debate rather than, you know, do some research. IQ being used as a metric is another one, since the results can vary wildly.
Quite clearly someone was sleeping or playing pocket tennis in history class.
As any of the great masters of propaganda of old can tell you in order for the news to be believable they need:
1. To fit the mental view of the world on the target. Example - the tripe fed to USSR cittizens about the west and the tripe fed to USA about the East - especially during the McCarthy era and Stalin rule. Another, more extreme example - the stuff which Albanians were conditioned to believe in and what the North Koreans are conditioned to believe in today. EVERY audience has its conditioning. This, by the way is the power of SCL, CA, etc - they allow you to tune the message based on the conditioning so it is more believable (*).
2. Mix of truth and current trending news with things that the subject audience will find plausible carefully tuned to fit 1.
There is VERY LITTLE DATA to discern the PROFESSIONAL fake message designed to influence an opinion. This is not a data problem. In fact, presenting it as a data problem makes me extremely suspicious of the data mongerers wanting to retain the SCL/CA like business.
It will also only get worse. In days past like it or not we all got a common newsfeed. If we were lucky, it did not contain too much propaganda (depended where one lived).
This is no longer the case because 99.9% of the population (including most of el-reg commentariat) peruse only single news sources which are "attuned" to their viewpoint. Nobody follows "neutral" or god forbid "enemy" news sources any more.
(*)I am deliberately avoiding current affairs so that we do not go down Godwin's law 2.0 but the conditioning of audiences on both sides of the "story" is beyond obvious.
Quite clearly someone was sleeping or playing pocket tennis in history class.
"Those who fail to learn from history are doomed to repeat it.". Well, history isn't being taught in depth anymore in schools so if we look at where we are now, the destination is a foregone conclusion.
@VRH:
" EVERY audience has its conditioning" " because 99.9% of the population "
absolutely 100% agree. And as long as a mostly conditioned audience continues to feed on the assigned and attuned "NEWS!!!" feed, the condition can be referred to as an echo chamber.
"If we were lucky, it did not contain too much propaganda "
Like it or not, we've all been dosed with propaganda, and tons of it, sadly I'm afraid much of it starts in the family home, prior to school. I'm personally of the opinion that it is the ability to reason logically about the data one is fed that is the imperative to communicate to our offspring, since there is no available unbiased reporting of just about any event that has occurred in the last 20 years (at least -- I've only found one or two off the wall, out in the boondocks wild-card reporter types that have the wherewithal to get their unadulterated word out) on anything remotely related to mainstream media, social media, or hell, even word of mouth. (not injecting long sidebar about specific local event that I was too close to)
I can only hope that my offspring have sufficiently acute reasoning skills. Hell, I hope *I* have sufficient reasoning skills some days.....
Does the very simple transparent mathematical formula work? In the example of mortgages do banks go bust? The 2008 recession was probably caused by greed and not following the formula otherwise we would have mortgage right downs and banks going bust all the time. So my question is why if something works do you want to add complexity? I sometimes get the feeling that neural networks and machine learning are being added just for the name and the perceived intelligence it adds where really a very simple transparent mathematical formula is all that's needed.
Don't get me started on the causes of the 2008 financial crisis. [That was 40 years in the making.] What caught my eye and apropos of your comment:
"It's non-linear math, the data goes through a lot of transformations, and actually it's a learning system so the data is changing all the time, so you can't actually explain how the decision is made."
Being only a (not so) humble econometrician is that this CEO doesn't even listen to his own data scientists. Unless someone is hiding their work*, all the systems are determined using a forward linear system with one, or more, feedback mechanisms. They are quite deterministic. Feed them the same data, get the same weighted neural net out. Unless you insert jitter which I've not come across. Yet.
Where the problem lies is that you have to keep all the live/training data around to justify the model you are using in your business processes. So you have all that data subject to the GDPR, used to create a model, for which you must keep all that data subject to the GDPR to justify your model. Hello circular loop.
*- Back in the late '90's, word got out at the hospital at the time that I had become something of expert around experimental design, statistical/mathematical methods, models and scientific computing in every field that might make use of them. It was my new thang. Anyhow I developed an autonomous, non-linear neural net to forecast future case management by population and the diseases/conditions that will be presented. It was a number crunching monster, especially as it would dynamically play with what non-linear functions should be used at each neuron. Autonomous as it was only told the data type. It figured out its own neural topology. Non-linear as we are working with biomedical processes. Yeah, I was over at the medical school for a few of my courses on their take on scientific computing.
Start it, let it run a week to ten days, come back. Hopefully it ran to completion. There is no way in hell I could explain that puppy's decisions. Justify the elements in the process? Yeah. Otherwise? {Shrug} Anyway, I have a pretty good bullshit detector on this stuff.
So, I interviewed for a role answering to the c-suite's at a data processing company recently who held masses of 3rd parties customer confidential data in a cloud env to process it, and the overwhelming thing I came away with about the entire company from the various interviews is that they don't give a rats ass about the data itself or if they have a corporate trustworthy stance, and their entire focus all about protecting their ip algo from walking out the doors because they trust none of their minimum wage coders and data scientists. No I didn't take the role.
So to distil the recommendations down to plain speak, be trustworthy = write nice words on the website and pretend to care so that people will keep giving them not fake data because it ruins their business model. Not unlike Facebook are currently desperately trying to do.
Me? I have a fb account. Perhaps more than one. Happily peeing in the well to poison their data sets while at the same time maintaining placeholder accounts to stop someone else spaffing things on my behalf.
Before the truth has its (insert clothing item here depending on attributed quote) on. Then surely all we need to do is produce a coverage velocity algorithm. Determine the mean distribution speed of all "news" and immediately discard the upper percentile as "fake news" .
QED
Just some cunning selection based on a pool of genuine articles, all in the box called 'your newsfeed' or potentially any news site front page because it's all dynamically generated anyway.
The selection gives a person a window on the world, and this can be made consistent across a number of sites if they all use the same analytics services and/or advertisers. This consistency becomes a very strong method by which to reinforce or reduce someone's strength of feeling about pretty much anything other than perhaps marmite.
Was facebook the only company to try the 'mood influence' thing (which I hope people haven't forgotten)? Who else is doing it - that hasn't been outed yet?
.
p.s. I also really really hate the expression 'fake news', not least because it evokes the image of someone saying 'noos'.
You can't look at my Linkedin profile because when I retired I first progressively pruned my history, and then after a few years when I was sure that nobody who might want to contact me didn't have my email address, deleted it. Of course I have no idea whether it's really deleted.
In the very early years Linkedin was useful but Gresham's Law applied. We now have Gresham's Law for news of all kinds.
The thing is, there are very few genuine investigative reporters left because nobody wants to fund them. Robert Fisk is now 71. Carl Bernstein is 74. In one of his books Robert Heinlein came up with the idea of a profession of fair, unbiased witnesses. Perhaps that's what we need, but how do we go about creating them?
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
