Facebook is Rating Users Based On Their 'Trustworthiness' (engadget.com)

Facebook has begun to assign its users a reputation score, predicting their trustworthiness on a scale from zero to 1. From a report: Facebook hasn't been shy about rating the trustworthiness of news outlets, but it's now applying that thinking to users as well. The company's Tessa Lyons has revealed to the Washington Post that it's starting to assign users reputation scores on a zero-to-one scale. The system is meant to help Facebook's fight against fake news by flagging people who routinely make false claims against news outlets, whether it's due to an ideological disagreement or a personal grudge. This isn't the only way Facebook gauges credibility, according to Lyons -- it's just one of thousands of behavior markers Facebook is using. The problem: much of how this works is a mystery. Facebook wouldn't say exactly how it calculates scores, who gets these scores and how other factors contributed to a person's trustworthiness.

Comments

Truth is not truth...

[The Original CDR]

Trustworthiness is the new truthiness?

Re:Truth is not truth...

[presidenteloco]

No. Some facts and theories are testable by the scientific community using scientific method.
Some are demonstrable by scientific tests simple enough for anyone (with sufficient resources) to reproduce.

Correspondence (of propositions and theories and terms) or not to measurable aspects of physical reality is a testable thing. Enlightened humans discovered that about 400 years ago.

Maybe you didn't get the memo.

Re: Truth is not truth...

[Anonymous Coward]

No scientists believe that global warming is 100% the fault of human activity. There is robust debate about how much is caused by humans, from fractions to majority.

You just demonstrated exactly the type of truth manipulation that you were trying to argue against. Only fanatics are entirely certain of "fact" without debate or room for doubt.

Re:Truth is not truth...

[fish_in_the_c]

sure, but how many of those facts are 'interesting'.

For instance. Let's just suppose one could prove beyond any reasonable doubt that withing 500 years the greenhouse effect would destroy the earth and make in uninhabitable by mankind.

( let's ignore the difficulty of proving that for the sake of the demonstration.)

You will notice what has NOT been proved.
a) that there is anything we SHOULD do about
b) that there is anything we CAN do about

why, because material science can't prove or disprove a moral proposition.

So while science can prove useful facts like. IF you do this ,you have a high likelihood of accomplishing that.
It is entirely useless when it comes to the first part of the preposition. That is to say 'should you do the IF'.

Re:Truth is not truth...

[presidenteloco]

"because material science can't prove or disprove a moral proposition."

Just for the sake of fun argument: One could posit an overarching moral principle which is possibly able to be scientifically and mathematically investigated, and a somewhat objective assessment of ranking of moral states might be based on that:

Here's one candidate general moral principle:
TLDR: Maximize quality complex-life-years summed over some spacetime region (set of situations).
                      Measured in bit-seconds per cubic-parsec (or cubic metre) of life-information conservation, perhaps?

You may recognize this as a vast but defensible generalization of "Thou shalt not kill" or the related "don't covet thy neighbour's wife", "don't steal" or otherwise mess up life-conserving societal cohesion. The Dalai Lama's "My religion is kindness" is also covered, as kindness promotes co-operation, minimization of energy-squandering and life-squandering conflict, and more energy efficient co-operative processes for keeping more people alive longer. It also handles the more controversial "women and children to the lifeboats first" and at least lets you quantify trolley problems. :-) It also covers bio-diverse eco-system conservation as a moral goal.

Details:
Each individual lifeform is most generally characterized by representing it as an information pattern that describes the complex form of the matter-energy pattern that is the lifeform, or perhaps alternatively by an information pattern that encodes the construction of the lifeform (the genome).
We will say that each lifeform embodies its characteristic information pattern. The stable, descriptive,complex information pattern is carried (present) in the form and genome of the organism, even if nowhere else.

The amazing thing about life is that it conserves locally many, very similar but not identical, copies of these complex-information patterns. That, most essentially IS what life is and does. A local particular-information conserving process. A negentropy machine working in an open thermodynamic system.

The complexity of the lifeform is well measured in terms of the bit-length either of its generative genome information-pattern, or if you prefer, the bit-length of the matter-energy-pattern-descriptive information-pattern. Normalized of course by near-maximally compressing the information in the information pattern (so that is close to a Kolmogorov-random bitstring).

The thing about long PARTICULAR bitstrings that can encode complex form is we do not expect them (the bitstrings) to be conserved throughout time in most physical regimes. Too much entropy or free-energy going on all around them. Conservation of the information is a tough job, achievable only through the form and continued functioning of the complex lifeform information-containers. Lifeforms carry forward the complex particular information through time, unexpectedly compared to the average rate of entropy generation in their region. That's life, precisely.

So I'm positing that morality was developed by humans to make life easier (more survival probability per unit of energy expended) for more people. A self-aware, environment-aware, intelligent-agent piece of life (humans) seeking to conserve life-iness through inculcated behaviour guidance. Originally just applied to people, but the extension to lifeforms/ecosystems in general, ranked by complexity, is straightforward (since let's not forget, we're inextricably part of that ecosphere.)

Of course there would still be room for argument, mostly over how to scope the situation-set, and then also about how to define life-quality (Zen and the Art of Motorcycle Maintenance?.) But the basic framework for morality assessment is there.

As long as you accept that my universal moral principle is a valid generalization of most if not all other common moral principles.

Re:Truth is not truth...

[presidenteloco]

No. I don't think you have understood what I said.

One person (and their live-long-and-prosper status) is worth a lot in my math, with that person being a bundle of a lot of life-complexity.

However, when it comes to: Given no other choice of a win-win scenario, do you knock the terrorist off the ledge to spare the 10 people they are threatening to blow up, well you can see how my principle would say, yes, kill the terrorist. So I violated "thou shalt not kill" but in order to "not allow to be killed" 10 other more life-affirming/supporting people. I stand by that math. Of course, my math also says that if you have an option that saves the terrorist and the victims, prefer that one. I frankly don't see how all of that is "unchristian". It's more about how you modify an overly simplistic rule when faced with complex real situations with multiple factors to weigh.

And sorry, a creed that says that a single individual is infinitely valuable is ridiculous and dangerous, when it comes to practical application. You are immediately drawn into "yeah but what do I do when faced with choices to help this infinitely valuable individual survive versus that one, like for example who gets the heart transplant." Your principle does not help. Since your principle is simplistic (for teachability reasons, no doubt), individuals trying to apply it must violate it using their own discretion, to handle real situations, without guidance for how to gradually modify the principle to fit reality. Not as helpful as the more universal principle.

And more importantly, my principle also supports Vulcan morality: The needs of the many outweigh the needs of the few, or the one.

Live long and prosper.

Zuckerbook == China?

[Rick Schumann]

Gee whiz Zuckerbook, you're starting to sound an awful lot like living under the communist Chinese government, aren't you?

Re:Zuckerbook == China?

[Rick Schumann]

Not so specifically as you're stating it, no, but I do remember the vast majority of people having been 'indoctrinated' that way by social media, to believe that anyone who wanted to preserve their privacy 'must have something to hide' and therefore must be criminals, terrorists, and/or pedophiles. I never fell for any of that, and as the pressure to be brainwashed by social media increased, my aversion to social media increased proportionately, and I don't use ANY social media anymore, and haven't for a long, long time now, and encourage everyone I can to dump Facebook, Twitter, and any other so-called 'social media' and (shocking!) actually be social with live people away from the Internet.

Re:Zuckerbook == China?

[swillden]

Facebook is an opt-in service not mandated by anyone or anything. Don't want to use it, then don't. The only time you "need" Facebook is if you're dealing with public relations, and even then, you're not using Facebook for socialising but for promoting.

A Chinese citizen does not have the same freedoms. Facebook and China are not comparable in any way.

I think the idea that this is going to happen is laughable, but it's worth pointing out that if the Facebook trustworthiness score were to be used like the Chinese "Social Credit" score, it wouldn't be necessary to force people to use it. All that would be required is for various entities to begin relying on the score to make decisions about whether or not to trust someone. In such a world, someone who refuses to use Facebook would be distrusted by anyone who relies on the trust score. If you wanted to be trusted, you'd need to participate, much the way that if you want to be able to borrow money to buy a house you generally need to build a history of borrowing and repayment on smaller loans first.

Facebook

[beep54]

Who is going to rate Facebook's 'trustworthiness'? Anyone....anyone??

Re:Facebook

[SeaFox]

The stock market? That is about the only thing that can make a difference.

Yup, if there's one thing we can depend on as a moral compass, it's the actions of large corporations. /s

I think it just might work.

[Gravis Zero]

Facebook is basically trying to define truth by using user input. Obviously this can be manipulated but I think by adding their own input to week out untrustworthy users, they just might be able to pull it off. There are a lot of variables that can be factored in to deem someone trustworthy but many factors can also be gamed. However, with additional input on who is a bad actor then it also discredits the users that trusted that user.

Russia is definitely going to fuck with this system but I don't think it will be a case of cat and mouse because there is a continuing basis for trustworthiness. It's just crazy enough to work.

The real news here is that Facebook may actually start doing some good instead of being a total parasite!

I know how this works

[SuperKendall]

Mention Infowars (even in jest) - rep score 0 forever until the end of you or Facebook.

Post link to Huffington Post article - A++++ GOLD STAR WOULD ALLOW TO POST AGAIN.

I posted just one political comment on Facebook once, expressing a desire that those on the left and right should talk to each other and not shut people out - so I'm pretty sure my trust score is like -5 out of 0-1.

Re:I know how this works

[omnichad]

I know you're joking, but even basing it on posting a link is a bad move. I may post a link to an article that is 100% wrong because I want to comment on how wrong it is.

Re:I know how this works

[Anonymous Coward]

I know plenty of democrats who behave that way but very few of us who are actually on the left.

The problem from our point of view is the right (democrats) and the far right (republicans) dominating all conversation, controlling all major media, and keeping out rivals to themselves despite most of the country really not being either of those. They refuse to talk policy and only want to argue with one another about trivia (Russia, Hilary's emails, etc.) instead of anything that matters while the country is suffering. Presumably because the donor class benefits and the politicians care most about the hands that feed them. We don't need neutrality. We need politicians who talk policy and vote for the interests of their constituents.

Get used to mystery

[presidenteloco]

"The problem: much of how this works is a mystery"

AI algorithms and knowledge-bases/trained models are already too complicated in their function for most people to understand. And they will get even more obscure and indirect in future versions, most likely.

Just as you don't know how I reached a decision or assessment, you won't be able to know how an AI reached a decision or assessment. We are just going to have to get used to that.

The chances are very high that the AI way of assessing will be more objective and principled, going forward, than most individuals' way of assessing.
If you like, to make people less suspicious, perhaps a convention of publishing the code and data in the assessment system (anonymized when references personal data) might be established. But how will this help? A few experts would be able to check it and vouch for its reasoning integrity, but nobody seems to believe experts these days since many of them seem "bought" anyway.

Black Mirror and 1984 all in one

[Anonymous Coward]

This is LITERALLY a "Black Mirror" episode on netflix, where people's entire self-worth is garnered by their social media score. This is big gov't censorship at it's worst (albiet Facebook is "sort of" a private company). Add that to the fact Facebook is incredibly biased towards the left, censors conservatives and actual truth at all turns, this is literally the end of free speech and the beginning of the "1984" book/movie. As another poster pointed out, link, post or like anything to do with InfoWars or pro-Trump, and your trustworthiness will instantly plummet based on their completely hidden "algorithms". Wow

Re:Conservatives vs Liberals

[ScentCone]

You're confused. You're thinking that the handful of idiots who walk around with tiki torches and fantasize about a long-dead political organization are actual Nazis in any way that actually matters. They're universally scorned and/or laughed at, and on the rare occasions they actually break the law, the get busted and prosecuted. What you're thinking of are the modern day brown shirts - the actual foot soldiers of totalitarian fascism. That would be antifa. Black uniforms, masks, weapons at "peaceful" demonstrations, spoken desire to kill political opponents and leaders, destruction of property and beating people bloody when there's any sense that someone, somewhere may not agree with their violent world view. And of course, they generally do NOT get in any legal trouble for bringing weapons into no-weapons-allowed areas and hurting people. They're all about intimidation, silencing people, and terrorizing those they don't like. And they're cheered on by the statists who now run the left side of politics in the US. There's your contemporary Nazis.