Catch up on stories from the past week (and beyond) at the Slashdot story archive

 



Forgot your password?
typodupeerror
×
Social Networks News Technology

Know-It-All Robot Shuts Down Dubious Family Texts (wsj.com) 160

Biting your tongue at yet another questionable article shared in your message group? Add artificial-intelligence fact-checker Meiyu, she will jump in with 'False.' From a report: The artificial-intelligence bot will interject in real time when she detects posts about the news, pointing out factual errors and alternative interpretations. The technology, created by Taiwanese developers, is a step ahead of most fact-checking apps, including versions offered in Brazil and Indonesia, which don't jump into conversations. Other popular fact-checkers, such as Snopes in the U.S., are public databases that users consult for reviews of news items. Meiyu quickly became hot in Taiwan, which had just gone through divisive local elections and is rife with rumors of China's interference in social media. The bot now has more than 110,000 users on the Japanese messaging app Line, which covers about 90% of the mobile-messaging market in Taiwan.
This discussion has been archived. No new comments can be posted.

Know-It-All Robot Shuts Down Dubious Family Texts

Comments Filter:
  • by Anonymous Coward

    Awesome, do it now. The criminal fraud Presidency needs corrected.

    • Re: (Score:2, Insightful)

      by lgw ( 121541 )

      Awesome, do it now. The criminal fraud Presidency needs corrected.

      And this is of course the problem with bots like this. The "facts" will always conform to a particular political world view - that's just how humans work. Imagine your feed where a bot installed by a big Trump fan labeled stuff as "fake news" whenever it disagreed with Trump. And that's exactly how it works in China. It isn't some hypothetical situation here: government control of "fact checking" is the very heart of modern totalitarianism.

      People have these naive ideas about "facts" being objective and b

      • Re: (Score:1, Informative)

        by Anonymous Coward

        So you're saying Trump hasn't lied ~8700 times since taking office, provably, in verified print or on camera? Or are you saying your head up Putin's ass is more important than the truth? Either way, kill yourself traitor.

        Trump hangs for treason either way, and his and your lies will be corrected whether you like it or not.

      • by Megol ( 3135005 )

        So a Taiwanese program used in Taiwan is controlled by China? I can't check the article myself as I have no login on the linked site and will not get one.

        • So a Taiwanese program used in Taiwan is controlled by China? I can't check the article myself as I have no login on the linked site and will not get one.

          That's utterly missing the point of his comment. Unless you think Taiwan is staffed solely by angels.

        • "So a Taiwanese program used in Taiwan is controlled by China?"

          If the company is run or staffed with KMT supporters then it might as well be controlled by Communist China.

        • by lgw ( 121541 )

          "Fact checking" is, in general, a process that always spirals into ideological control and enforced orthodoxy.

      • Re: (Score:3, Insightful)

        by Shotgun ( 30919 )

        Worse than that, some "facts" are impossible to establish.

        We all now know that Jussie Smollett is a liar, but is Ilham Omar really an antisemite? That is, was her apology sincere?
        We know that Trump publicly asked Russia to find Hillary's emails, but did he know if they had them, or was he sarcastically playing on the news of the day?

        Most people's version of "truth" will depend on how they interpret vague things that are essentially unknowable.

        • To be fair we don't *know* that Jussie Smollett is a liar. He still stands by his story and the evidence hasn't been presented at trial yet.
          Right TruthBot?
          That is correct.
          • That's a perfectly straightforward solution. For claims that can't be established as true or not, simply label them as inconclusive. For complex subjects, enumerate the claims, and evaluate them independently. Jussie Smollett might have been proven to lie about things in the past, and he might have been proven to tell the truth in the past as well, but those have minimal impact on the current issue, whose truth has not yet been determined.

          • Don't need a trial for the court of public opinion. Just ask yourself, is Russian Collusion proven in court yet?

            Yet the media and sycophants in congress all keep spewing that story. Or are we gonna need to quote Nazi Propagandists about repeating lies till they become truth ?

        • Worse than that, some "facts" are impossible to establish.

          That is true. In fact, the truthiness algorithms probably have some calculated level of confidence for all inputs. The practical use of the algorithm doesn't require raising an alarm for every input that is determined to have some level of untruthiness. Rather, the alarm can be raised for just a high level of untruthiness. Furthermore, since it's the oft repeated untruths that do the most harm on a societal basis, the alarms can be further restricted to those high-level of untruthiness statements that a

        • by Tom ( 822 )

          Truth is never subjective, but our knowledge of it can be limited or false.

          If someone said something, that's a truth. He said it. If we have recordings of it, we can prove that he said it (ignoring forgery for a moment). Thus the truth value of a statement that says "he never said ..." can be established and evidence can be provided.

          A machine like this does need to know the "undetermined" answer. It cannot work binary. "I don't know" has to be one of its answers as well as "there is conflicting evidence, he

      • All of the above is why any fact check database like this needs to have transparency, crowd sourcing, error bars, reputation scores, and metamoderation.
      • The white man came up with the idea there is "objective truth" and used it to oppress peoples of color. The Enlightenment's ontology, rooted in the new science of the 17th century, created a vision of human beings in nature which provided weapons to a new race-based ideology which would have been impossible without the Enlightenment.

        The entire idea behind today's Left-wing thought is that there is no objective truth, only differing points of view, all equally valid. For example, there is no valid genetic b

        • I have never read a more misinformed bigoted post that was trying to be honest.

          The fact is, IQ (Intelligence Tests) are so well studied that they have all but eliminated cultural biases, and are generally applicable to everyone world wide.

          In trying to dispel stereotypes, you only succeeded in making a fucking huge one.

          The only thing we can do with group identity and genetics is to show that humans fall within varied distributions that roughly occur along a bell shape curve. Most people, in most groups, over

          • Racism dressed up in progressive clothing is just as ugly as the KKK and Nazis.

            That's sig-worthy, bumpersticker-worthy, all kinds of worthy.

      • ideas about "facts" being objective and bias free, and that's great, but fact-checkers never are

        Okay, I will agree with you on this point at least. But... so what? From a practical standpoint your quibble seems to be entirely useless. Regardless of the bias of the fact checkers, facts still exist and falsehoods still exist and so facts need to be checked.

        We have a bit of additional protection in that, unlike other journalists, fact-checkers' goal and purpose is to suppress bias as much as possible. So they have some motivation, more so than other journalists, to keep their information as neutral as

        • by lgw ( 121541 )

          We have a bit of additional protection in that, unlike other journalists, fact-checkers' goal and purpose is to suppress bias as much as possible.

          All the major fact-checkers in the US today have descended into politics pretty far already. Some are blatant political propaganda.

          And do you think people want neutral fact checkers? Of course not. They want their beliefs confirmed, and will find fact checkers that do that [faithfacts.org].

          • Yes, fact-checkers are subject to all of the standard pitfalls of journalism. As well as all of the standard solutions to those pitfalls. The question remains: what is your point? Where are you going with this? Are you seriously suggesting that we should just throw up our hands and give up on the concept of truth? Facts still need to be checked.
            • by lgw ( 121541 )

              My point is that the false authority of "fact checkers" is a tool that really serves no other purpose than abuse. It is a tool of ideological propaganda, and should be thusly despised.

              Do your own research, or "follow" someone who does. Either way, you can only find the truth at the end of your own journey of discovery.

              • or "follow" someone who does

                What could we call such a person, I wonder... I'm going to go with "truth verifier."

                You've made a giant leap from, "fallible humans who don't always succeed at overcoming their biases," to, "false authorities worthy only of derision and hate." I don't see how that leap is in any way justified, and it's impractical regardless. We're always going to need fact checkers, even if you don't want to call them fact checkers.

                I read an article a while back in a philosophy publication in which a guy made the sug

                • by lgw ( 121541 )

                  There's a significant difference between trusting someone you know, and trusting an abstract "fact checking service" with an unknown agenda. Much like movie reviewers - most are useless, but sometimes you find one whose biases you understand (but even then, sometimes you discover they've been taking bribes), and sometimes you just go with the opinion of a fried who got there first.

                  The last bit is how I work, for the most part. I benefit from the professional experience of my friends, when possible, or und

                  • You're reading a news site. If the only facts you're willing to accept are those which are directly connected to you friends' employment, or which you research extensively yourself (Whatever that means. I have no idea how you research something if you're unwilling to accept any information which isn't connected directly to one of your friends.) then... why are you here? What could Slashdot possibly offer you, if you're unwilling to believe anything that anyone here says?

                    You do say that you're willing to
                    • by lgw ( 121541 )

                      Plenty of stuff is interesting even if you can't be sure whether it's true. The political stuff on Slashdot is worthless except as debate practice, but there's still plenty of technical/scientific stuff. I've learned a lot over the years from looking into "shit some guy on the internet said". When you get away from politics and economics, it's much easier to verify stuff and decide who has earned your trust.

                      E.g., I completely flipped my opinion on dark matter and dark evergy thanks to some smart-sounding

    • Re: (Score:1, Troll)

      by Tablizer ( 95088 )

      Predicted response: "Fake bot, totally rigged, believe me! I know more than these bots, I'm a technical expert. They merely echo lyin' CNN, any loser can do that. Build a bigly firewall around those yapping rusted toasters that nobody listens to. Lowest ratings of all machines ever. My desk's diet Coke button has better ratings."

      • You forgot Cheeto and Oompah Loompah references.

      • Their suspicion of the bots being rigged has some merit. The EU runs a small bureau to (manually) combat fake news in a similar manner, but it has been caught out for being wrong about several news items. And it looks like it wasn’t a case of honest mistakes, but politically motivated spin.

        These bots can be an awesome tool to achieve the same on a grand scale. Simply teach them to sniff out discussions about truths you want suppressed, and have the bots chime in with carefully crafted official nar
  • Re: (Score:2, Interesting)

    Comment removed based on user account deletion
    • by Megol ( 3135005 )

      I think you'd be disappointed. Someone that thinks scientific consensus is a peoples vote...

      • Continental Drift called, it wants you to be less credulous of the Divinity of the extant process.

      • Consensus isn't science. People claiming it is don't understand what science is and that it isn't a popular vote. ;)

        • by gweihir ( 88907 )

          Scientific consensus is science. It means a lot of smart people have looked and did a) find convincing evidence for it and b) failed to find any reasonable contradictory evidence. As a result, scientific consensus is nothing like ordinary consensus. This faulty reasoning is however a favorite of anti-science people that do not like the facts that science finds and think that if they just believe hard enough in their own fantasies, reality changes.

          • by Tom ( 822 )

            This faulty reasoning is however a favorite of anti-science people

            I'm not sure anymore. I think they intentionally ignore the science for their pet topics while perfectly ok with it in other topics. It's like the preacher telling about modern technology being the devils work - in a church built with modern building techniques over a computerized speaker system. Or the islamists burning evil western electronics and filming it with smartphones.

            There people aren't necessarily "anti-science". They are just in their specific field of interest so dead set on their preconception

            • by gweihir ( 88907 )

              Well, I regard people that selectively accept only the science that fits what they believe to be fundamentally "anti-science". Science is not a bazaar where you select what you like and ignore the rest. You can either be rational and accept the full package (whit all the sometimes very uncomfortable facts that brings) or be irrational and treat scientifically well-established facts as things that are up to discussion by laypeople or even as things that can be conveniently ignored. Of course, Science has als

              • by Tom ( 822 )

                But it is emphatically not up to non-experts to determine which is which.

                This, btw., is a problem - science isn't democratic.

                We'll have to solve that problem if we want both of them to survive.

                • by gweihir ( 88907 )

                  I agree that it is a problem. But it is that Science is not egalitarian. If you want to actually have a voice in Science, you need to devote a larger part of your life to it and you need some pretty specific talents to start with. And that voice needs to come with proof for the statements made, so just pushing opinions (all that most people can actually do) is not going to cut it. Hence most people cannot get a voice in science.

                  That said, the problem is made worse by regrettably quite a few scientists that

                  • by Tom ( 822 )

                    But it is that Science is not egalitarian.

                    In theory, yes. However, even scientists are only experts in their own narrow field and have to trust the other scientists for everything else. And science and its results touch you literally everywhere. From air pollution to the food you eat to the tech you use and increasingly (sociology, psychology) the politics, society, culture and personal interactions of your life.

                    We already see people abusing pseudo-scientific arguments in political ways (not just in politics, also in pushing agendas outside of part

                    • by gweihir ( 88907 )

                      But it is that Science is not egalitarian.

                      In theory, yes. However, even scientists are only experts in their own narrow field and have to trust the other scientists for everything else.

                      Actually, no. First, any real scientist does understand the scientific process and what it can and cannot deliver. And second, a scientist is able to verify claims in a different field to a much deeper degree than a non-scientist. Sure, the quality of that verification drops the more removed the other field is, but it is still much better than what a non-scientist can do. Hence at least good scientists are a lot harder to fool and at the same time do understand a lot better when they are incompetent with re

                    • by Tom ( 822 )

                      Actually, no. First, any real scientist does understand the scientific process and what it can and cannot deliver.

                      Absolutely. If you are a scientist or knowledgeable in the scientific process, you can easily spot obvious bad studies and judge if the approach of a study is sound even if you don't know the specific field. But you are still far from being able to decide the truth value of a hypothesis. You have to rely on the peer review process.

                      (on ethics)

                      I see that our definitions of the word differ. For me, ethics is more personal while you consider things that I would see in the field of politics or sociology into it.

                    • by gweihir ( 88907 )

                      Agree on your statement on Ethics. There are different ways to see it. As to Nuclear, Fucku(p)shima made the "but in the West" argument obsolete, because it does not get any more technologically advanced than Japan. If they cannot do it safely, nobody can and it is really not a tech problem.

                      Anyways, it has been a while since somebody on /. gave me a good discussion, thanks!

                    • by Tom ( 822 )

                      As to Nuclear, Fucku(p)shima made the "but in the West" argument obsolete, because it does not get any more technologically advanced than Japan.

                      Though the argument has been made that Fukushima proves that a western-built reactor won't blow up and spray radioactive clouds over an entire continent.

                      Anyways, it has been a while since somebody on /. gave me a good discussion, thanks!

                      Same to you. It has become rare in the last few years.

    • by Ichijo ( 607641 )

      That detects people asserting logical fallacies and tells them that they don't know what they're talking about.

      And I want one that detects people asserting that logical fallacies means people don't know what they're talking about, and tells them that they don't know what they're talking about. (Argument from fallacy is also a fallacy.)

      But not one that tells mine that it doesn't know what it's talking about, because that would be just crazy [imdb.com]!

      • Not all logical fallacies are false, simply because they are a fallacy.

        Slippery Slope is a logical fallacy, not because it isn't sometimes true, because it is sometimes true. Slippery slope is a logical fallacy because one cannot assert it as a truth, because it might be false. Dismissing a Slippery slope argument as "false" is also a logical fallacy, because it can be true. The point is, it isn't definitive either way, which also makes it a logical fallacy.

    • Re:I want one... (Score:5, Interesting)

      by Aighearach ( 97333 ) on Friday March 01, 2019 @02:44PM (#58200978)

      Scientific Consensus would be argumentum ad popolum if it was true that consensus determines facts, however, it does not. It merely flags particular theories as known-to-be-accepted-in-the-mainstream.

      Science is a process, not a set of theories or conclusions. Therefore, you identified the correct fallacy in the argument, but in the wrong part of the argument. Where they're engaging in argumentum ad popolum is when they're saying that because there is a "scientific consensus," therefore the body of the consensus is also correct, or worse, sometimes they'll even want to call it "proven." But science does not and can not attempt to "prove" things. Science is merely a process by which you can attempt to repeat things; the personal goal of scientists is often to discover the "why" of things, but science doesn't actually do that. And the good ones know it! Their personal goal of understanding "why" is separate from their professional goals of taking additional steps in a process. But conclusions are not one of the steps that is even part of the process, so consensus about what the current-best-answer is doesn't even touch on or support coming to conclusions about it!

      "I think, therefore I am" is a complete failure, but because the lesson is valuable (if I don't exist it doesn't matter if I'm wrong, if I do exist but believe I don't I won't be motivated to survive, therefore belief in existence is the only answer with potential utility) we also value scientific consensus; not because it is ever correct, or capable of being correct, but because it supports additional attempts at understanding.

      • Re: (Score:1, Interesting)

        Scientific Consensus isn't either. Science isn't based on consensus, and consensus isn't always based on science. There are a lot of butthurt feelings whenever I cover this exact point, almost like feeling matter more than facts.

        Consensus is often a blockage that prevents additional understanding, because the new understanding might have difficulty because it is contrary to the current consensus. Real Science tries to avoid consensus, largely because it has been wrong so often in the past.

        The weird thing is

        • I would say that on average the scientific consensus is a better approximation or model of the capital-T Truth than the average other explanations that exist. That's a little bit meaningless, but the idea is is that the current consensus replaced a previous consensus that wasn't as close to the true explanation or understanding of the universe. New ideas can come out and change the consensus, but the consensus exists for a reason. It's not that people were wrong or idiots, but often just do to the limitatio
          • consensus can be limiting because it naturally opposes new radical ideas. However, most radical new ideas are complete shit

            This is the problem, in a nutshell. E=MC**2 was "radical" idea. Simplistic, yet radical. Quantum Mechanics was a "Radical" idea. It still is goofy weird ... but ... for now ... it appears more or less correct. But I have no doubt as we further discover new and improved methods, this too will be replaced with something better.

        • Real Science tries to avoid consensus, largely because it has been wrong so often in the past.

          Wut??

          You honestly believe that? Talk about "feelings matter more than facts"! First, the number of things mainstream scientific consensus has been regularly shown to be right about vastly dwarfs the number of times it's been wrong - which is why it's still so important to the scientific process (and if you really haven't noticed that about virtually every major field then you truly are deep in denial).

          Second, on the occasions it's been wrong - a new, more correct consensus formed as soon as there was compel

    • by gweihir ( 88907 )

      And fail. Scientific consensus is not "argumentum ad popolum", as to be part of it requires some actual qualification and insight into the topic. Your "argument" (which is actually a fallacy) is however popular will all those that do not like what science finds out and would much rather regard their own fantasies as the truth.

    • when they get told that "scientific consensus" is a euphemism for argumentum ad popolum.

      It isn't. It is a euphemism for "argumentum ab auctoritate" i.e. appeal to authority. The point is not that lots of people believe it but that experts in the relevant field (who are presumed to know the truth) believe it.

  • by Anonymous Coward

    From the "free speech" advocates who go all fluttering concern when the idea of correcting false speech and alternative facts is broached.

    • by Shotgun ( 30919 )

      Allowing me to add a feature that will automate fact-checking with a source I trust is one thing. Having Facebook, Google or Twitter secretly do it is another ball of worms.

      • Facebook, google and twitter as fed by Russian and Chines trolls.
        usually the real problem is establishing trust of the sources. Much of the divide we have in politics is because 'traditional' media has proved time and time again it entirely biased to a particular political and social agenda. Most alternative sources don't even try to hide their bias, so if you want to guess at the truth you need to read between the lines.

        • Re: (Score:3, Interesting)

          The problem with the MSM is that it repeatedly sides with "narratives" (story telling) and not with just telling us what happened. It is the story telling (fibbing) that gets them into trouble.

          All one has to do is look at the two recent cases(Covington Catholic, Smollett), where the "narrative" that was being told didn't actually happen, and certainly not the way it was being reported initially in the MSM. In both cases the "narrative" was the story, not the actual events.

          People want to understand why "Fake

    • From the "free speech" advocates who go all fluttering concern when the idea of correcting false speech and alternative facts is broached.

      'Making sure things are correct' isn't the problem. The concern is the implementation:

      1. send text to fact checking group.
      2. parse text as to whether it contains information requiring response.
      3. formulate response and decide whether it needs to be sent.
      4. actually send the response.
      5. (probably) keep copy of text for aggregate data mining.
      6 (probably) GOTO 1 until 2 = FALSE.

      This is literally the infrastructure required to correct 'wrongthink'. While sure, we might agree with the immediate implementation of

  • Now angry RTFM neck-beards will be out of jobs.

  • From another recent post [slashdot.org]: "Social credit offenses range from not paying individual taxes or fines to spreading false information..." [emphasis added]. That would make for very swift "justice" indeed. Make a wonky claim in a "private" chat group while you're on the bus, and get thrown off immediately. No soup for you.
  • What a crock of sh*t. (As a side note, why does sh*t come in crocks?)

    • by Anonymous Coward

      The "chamber pot' is an ancient device, pre-greek. It's a, generally ceramic, bowl that, before plumbing, you pissed in rather than go outside at night. Shitting in the chamber pot was socially unacceptable, except for children with diarrhea. Hence the term "crock of shit".

  • Awesome (Score:4, Funny)

    by Locke2005 ( 849178 ) on Friday March 01, 2019 @03:16PM (#58201232)
    What happens when you send it a continuous stream of Trump's tweets and speeches?
    • by Anonymous Coward

      That was probably its 'false' training set.

  • Any chance this bot is sarcasm-impaired?
  • These things can only do what they are trained to do.

    I can see it now:

    AI: False! Human babies are not human until you take them home from the hospital. {Source: dem lawmaker}

    AI:False! You did get to keep your doctor, and your premium went down! {Source: who knows}

    Etc. ...

  • and maintain it's all just biased opinion and faith,
    please feel free to jump off the nearest cliff.
    I caution you against this.
    My unsubstantiated faith-based opinion (just as good or bad as anyone else's about anything)
    is that gravity will accelerate you rapidly toward the surface below and that you will die from the sudden deceleration after that.
    But what do I know?

    Seriously though, it's getting to where I refuse to have an extended conversation about a serious topic with anyone who cannot demonstrate basic
  • should sufficiently confuse the bot (Liar paradox [wikipedia.org]).
  • All sorts of surprises await the fools that buy lie-detecting robots! https://i.pinimg.com/originals... [pinimg.com]

  • by nospam007 ( 722110 ) * on Friday March 01, 2019 @07:00PM (#58202392)

    I'm married to such a bot.

  • This is actually a very, very interesting application for AI.

    I don't want it to judge. But I would love to have a Firefox plugin that scans texts and adds comments. So the next time someone posts "vaccines cause autism", there'll be a small symbol and you can click it to expand to a short summary of facts. Just the facts, e.g. "99% of scientific studies deny this claim, on the Internet the opinion is largely found on social media, 3 of its main proponents were sentenced in trials about the question, found t

Technology is dominated by those who manage what they do not understand.

Working...