“We live in an age of information, it is said again and again. But that doesn’t mean we live in an age of good information” ~ Rebecca Rosen
The above quote nicely sums up where we are right now. We need better ways of analyzing the veracity and integrity of the multitudes of information we meet with everyday in greater quantity. Skeptical readers perusing the Internet try, and often fail—not that it’s a bad thing, it only shows their human—to separate the good information from the bad information; the good science from the bad science; and the meaningful statistics from the meaningless statistics. This paradigm, of needing to verify and to fact-check everything, is going to change soon. Some time ago, I had the clever little thought—I don’t have many so I have to cherish them—that one day soon, someone will invent, or create, the Universal Fact Checker (UFC), most likely, as a browser plugin (an app for your browser that performs a task). Recently, something similar has been created, but I’ll get to explaining that shortly. First, I want to explain what I think the UFC will be. I envision it as an artificial intelligence (AI) that scours what you read informing you of dubious, false, or outdated claims, providing instant fact-checking on the spot—just as Fact Check does for US politicians; just like medicine does to snakes oilmen; and what science does to non-science. The key difference being is that it is with you at all times at the point of contact, as you absorb new information. You will not have to seek it out, or even to remember to seek it out, it will just be there karate-chopping bullshit in the face, like Penn & Teller, but, always there. Let’s face it, how many of us spot-check everything we learn? Not a single one of us. There simply isn’t enough time to do so even if you wanted to, and even if you had to. In such scenarios that we are in almost everyday, the logical solution is not to accept it as fate, but to invent a technology that alleviates the problem—inability to check and retain every piece of information provided to us—and performs the necessary tasks orders of magnitude better than we could.
How Might It Work?
Picture this: imagine you’re reading some pseudoscientist’s take on autism, intelligent design, theistic evolution, quantum healing, or whatever other woo you can shake a scientific stick at, but never makes it goes away (as not everyone will read it, or even have the scientific training to understand it) but, as you browse and absorb, your trusty little UFC scours ahead, subjecting every word, statistic, number, sentence, and paragraph on the page against empirical, peer-reviewed science and academic works highlighting the paragraphs that profess false certainties or provide dubious claims. In other words, MMA’ing the hell out of pseudoscience (I had to put a bad pun in somewhere). Only the strongest claims—evidence-based claims—will survive; what we would otherwise call good science; which is, what we would otherwise call—for lack of a better word—the truth.
Consider an example: (1) A website details the increase in autism rates in the last several decades (true). (2) It then goes on to say vaccines contain thimerosal (partially true). (3) It, then, continues on to say that since thimerosal contains the neurotoxin mercury (true), comes to the conclusion: (4) vaccines cause autism (false). So, how might the UFC access such a claim?
(1) The first section, after having been UFC-assessed, remains untouched because there really has been an uptick in autism rates. Though, if you happen to hover your mouse over it, you will be informed that much of the uptick has been due to a redefinition of autism, and, doctors becoming more aware of autism, thereby, increasingly diagnosing it instead of the condition going unseen or misdiagnosed. So, it is quite likely that the uptick in autism rates is not really an uptick at all, but merely, properly accounting for it for the first time, though still comparing it to the previous underestimated counts. (Of course, it will also tell you that it is a hypothesis, the leading hypothesis, but still not decidedly proven, yet far in advance of any other leading hypothesis.)
(2) The second section being somewhat factually based, is highlighted in orange. As a curious observer you, again, hover your mouse over the highlighted paragraph and a side-bar appears informing you that thimerosal was removed from vaccines by the summer of 2001, excepting the flu and tetanus shots. So, the statement, being as it is a generalization, has tried to lull you into a false certainty—and in this case, failed. You become slightly more suspicious of everything else the article professes to know.
(3) You move on to the third section, and notice that, it too, is highlighted in orange, with a sidebar informing you that methyl-mercury is a neurotoxin, but it (methyl-mercury) is not found in those few vaccines that still contain thimerosal (or any vaccine that ever contained thimerosal), as mercury in thimerosal is bound as an organic ethyl-mercury; it thereby being rendered impotent and easily filtered out by your kidneys, and, therefore, cannot be a neurotoxin. Your suspicions continue to increase.
(4) The fourth section, you’ve now noticed, is highlighted in red as the conclusion does not follow the logic deductively, but rather, inductively, and even then, in a series of inductive leaps with no evidential threads to support the leap from one to the next, so it’s closer to say that they are purely imaginative leaps. The sidebar will inform you that studies looking for any causal thread, which have cumulatively looked at millions of children, have found not even a simple correlative example between thimerosal (or vaccines in general) and autism, or any other disorder. It will tell you that in studies that looked at vaccinated and non-vaccinated kids, they have the same rates of autism, but overall, vaccinated kids get less sick. You now close the webpage and never visit the website again.
Now, wouldn’t that be a sight. Every creationist, anti-vaxxer, homeopathic, quantum healing, feng-shui, talking-to-the-dead website would be littered in orange and red paragraphs. The websites of the Thinking Mom’s Revolution; of Joe Mercola; of Natural News; of Age of Autism; of Reasonable Faith; of Answers in Genesis would become virtual ghost-towns, almost overnight (well, so the theory goes). They will cry foul, they will bitch, they will whine, and complain about being censored, and that it is all a conspiracy to keep the truth from you, because of course, only they have it. Some will listen, I hope most don’t. It will be true, their future babble about censorship, that is. But it will be censorship by good science, and since good science is what nature has regarded as true, it will be censorship by nature, or as I prefer to call it, the universe. (When people refer to nature, they refer to the insignificant speck of dust that is the Earth, but the Universe is where the action is at.) Michael Specter said it best in his book Denialism: “What can be understood and reliably repeated by experiment is what nature regarded as true.” Indeed, there will be a conspiracy, there will be censorship, but, it will be imposed by nature, and therein shall we find the truth.
I’ve meant to write this post for some months, but never got around to it. I finally did so after reading two interesting articles in close succession: one in The Atlantic by Rebecca Rosen, Is It Journalism, or Just a Repackaged Press Release? Here’s a Tool to Help You Find Out, and the other on the open-source science journal PLOS ONE titled, Text Mining Effectively Scores and Ranks the Literature for Improving Chemical-Gene-Disease Curation at the Comparative Toxicogenomics Database. (I highly recommend you read both before continuing, but if you don’t have time to read the articles, I will summarize—inadequately I might add, so read them.)
The first is a tool, named Churnalism, and it has been created to identify plagiarism in the media. It will allow users to submit or post articles and have the language checked against press releases, Fortune 500 companies, and government sources. This will help the would-be reader separate the wheat from the chaff, the original from the copied, and the reportage from the self-congratulatory, and subjective, press release. You’ll have front-row seats as the reportage, reporters, blogs, and online media without integrity fall to the wayside. In short, it is a simple way to instantly check the integrity of those whom we trust with reporting the truth. This tool has the potential to cull those with false pretenses. (You can even install it as a browser plugin so it automatically identifies those articles that have plagiarized. Just as I hope the UFC will do one day—hopefully soon.)
The second is yet another tool created to serve a specific need performing a different, though equally important, task (at least for scientists, though if it helps them, it helps us all). There are thousands of scientific studies being published everyday. (The open database, PubMed, alone publishes a new study every minute, and there is, perhaps, 50 million studies published somewhere.) No scientist can keep up with it, though it doesn’t stop them from trying. But, an inordinate amount of time is wasted weeding out non-relevant studies. If scientists could find a reliable way to accurately and quickly accomplish that task, it would, well, free up more time for them to do more science. So, a few scientists created a sophisticated algorithm that read through 15,000 papers going back to 1926 on metal toxicology and, using inputted indicators of article relevancy, novel data content, interaction yield rate, mean average precision, and biological and toxicological interpretability (you don’t need to know what these means) was able to, 85% of the time, rank the studies accurately in their relevance so that precious research time (and PHD students) could be focused towards those studies most conducive to their ends. Now, that is cool! (And so useful, but cool comes first.)
As I made the case earlier, this seems to be the beginnings of the left-hook out of left-field that pseudo-scientists will receive, and, hopefully, a lot sooner than many expect. These two programs, pieces of information technology, will not sit around unused and stagnant; others will take it, play with it, evolve it, and twist it to new purposes, and I hope one of those gifted folks turns it full-force towards the elimination of pseudoscience. Nothing is more relevant today than removing the influence of pseudoscientific jibber-jabber from the discourse we should be having on vaccines, nutrition, more importantly climate change and biotechnology, and perhaps even economics and politics. I can see no barriers to its implementation (aside from cost, which, as I’ll explain in a few paragraphs, is only a short-term problem).
I’m sure, by now, that most people know about IBM’s Watson beating two human opponents (the two best human opponents I might add) in Jeopardy; a game based on the nuance of human language. Watson, an AI, was able to deconstruct the language, understand grammar and syntax in the context of a question, and probabilistically match it to information it ascertained from Wikipedia. (That is, it wasn’t trained to play the game and had to figure out the answers all on its own in a similar manner to how our brains work.) Watch this video to see just how formidable Watson is (4-minutes long). You’ll even see most of the time that when Watson is beaten to the punch that he had the correct answer as well. Watson is now being trained as a medical assistant, and will be most instrumental in analyzing the totality of medical research and new studies coming out every day that a doctor could not hope to keep up with, and helping said doctor in correctly diagnosing patients reducing errors and cost, increasing health, and improving lives along the way. Watson, the fact checker, could be, in a few years, capable of the reasoning in our vaccine example above, if not already. And if IBM is this far, then other companies aren’t far behind. In fact, Ray Kurzweil, the futurist, is working to fully develop a personal, super-intelligent, and always online virtual assistant at Google that can read and understand the semantic content of the web at large. At that point, it will be possible that you’ll no longer have to search for stuff. You’ll just ask questions instead and empirically relevant, sound answers will be displayed. (Perhaps, this explains why Google is moving into hardware: Google Glass, self-driving cars, and the takeover of Motorola. No search results when you ask a question, but that is merely uninformed speculation.)
Instead of searching for when Neil Armstrong landed on the moon, where you may have several moon-landing-was-a-hoax results on the first page, you’ll only get the real, empirical answer: July 20th, 1969 alongside a photo of him, you know, actually standing on the moon.
Instead of spending long hours trying to research vaccine safety, where, as a non-medical professional, you can’t tell who’s giving you sound advice and who isn’t; and where, subsequently, a lot of good information is mixed, and lost in, a mountain of bad information, you’ll simply ask: “Is the DPT vaccine safe for my child?” The unambiguous answer will be yes, linking to the multitude of peer-reviewed studies (and only peer-reviewed) on the subject as well as, perhaps, explaining the pro’s and con’s of the quality of the studies, their methodology, any biases, statistical significance, and so forth. It will do this, perhaps, while also showing you the statistical advantage and risk-benefit analysis of not vaccinating your child, so that you may make your decision within the full context of available information bypassing your human heuristics that often ignores several important factors in valuing and acting on information.
Instead of having to filter through creationist babble about when and how the Universe was created, instead, you’ll ask “When and how did the Universe come into being?” The answer will be: “13.82 billion years ago. This data was ascertained with help from the Kepler and Hubble space telescopes, from WMAP, experiments in particle accelerators etc etc etc, and the best-supported hypothesis of creation at this time is a quantum energy fluctuation that instantiated itself into a system of net-energy zero that then forced negative space to expand to compensate for the positive energy instantiation, so that the system (Universe) remained at net zero energy.” (Of course, the super-intelligent machine will find a way to say this, or whatever the correct answer is, if it has changed or been refined, in a far more precise and succinct way than I have.)
But, where will these answers come from? From empirical, peer-reviewed research of course. From the hard and soft sciences, from academia, from open-source journals, and the avalanche of historic data just sitting around drawers waiting to be digitized, analyzed, and parsed through.
While the scenario I provided above—the autism example—is probably not going to happen for some years; for it takes an immense amount of computation and advanced algorithms. While these exist, they are supremely expensive, and considering that the UFC would be most useful as a free plugin—just as I have the churnalism plugin in my Chrome browser that automatically warns me if plagiarism is found—there is, as yet, no profit motive. (However, the profit motive is only necessary when the technologies are expensive. As they get cheaper, it will no longer be necessary.) But, because technology, particularly information technology (IT), is so awesome, it’s only a matter of time before it becomes cheap enough. (As IT’s price-performance per constant-dollar roughly doubles every year like clockwork. Therefore, in 10 years, the technology will be 1000 times more powerful while costing the same, adjusted for inflation, as now.) It’s only a matter of time before it is cheap enough.
Mark the calendar friends, Churnalism and the Science Text-Miner are only the first step. When the UFC arrives, it will come out of the gate swinging. At first, it will be simple, but it will iterate quickly and quicker until it encroaches upon, enveloping and suffocating, all the fields of pseudoscience, and real science will win. How could good science not win? It offers unlimited expansion, untold benefits, improves our lives in a very real way, and—again, for lack of a better word—has the good manners of being true. Pseudoscience appeals only to our vanity and ego and little more, it can only win in an environment where it is not selected against, such as the current (and past) environment where only a small percentage of the population are scientifically trained, but as soon as the tools of skepticism become available to one and all, it will be relegated to the dustbin of history, a future bedtime story told to kids who understand that having bad, non, or no science is as scary as the bogey monster now is to many… (If you doubt the sincerity of that statement, as I’m sure many will, then I invite you to move back to the Rift Valley in Africa and live without the benefits that observation, replication, and innovation have bought us, and which have resulted in the tools of our survival and eventual ascendancy. Those tools, which have bought us prolonged healthy life, increased food production, clean water, reduced infant and maternal mortality, and this webpage did not come easy. Billions worked, and died, for them so that we may be where we are now. See how long you last without shelter, tools, binoculars, night-vision, vaccines, weaponry, clothes, wheels, and, most importantly, fire.)
Impossible to say, but, it is only a matter of time. There is nothing forbidding it, our AI’s today are quite powerful, and information technology is getting cheaper predictably, every single year, so, it follows that our AI will only become more powerful, exponentially so. It is only a matter of time. When it does come, either next year, in five years, or in ten, hilarity will ensue, but more importantly, good science will finally and fully claim its status in the game of thrones played for with truth-claims for millennia Nothing will unseat it thereafter; well, nothing without a regress to the past. Lives will be improved and prosper; economies will grow and become more efficient; and, for good and all, better knowledge will have a selective advantage, and false knowledge will, for the first time in 200,000 long, agonizing, and painful years, have a selective disadvantage. Good riddance! The byproduct of our dear UFC will be, that, our minds will almost seem to perform as if on steroids. That is something I’d sign up for in an instant.
“Science is not a democratic process. Scientists don’t line up and say ‘gee,’ we really like this theory, let’s all vote for it. That’s not how it works. What we do in science is we find what explanations work.” ~ Eugenie C. Scott (Biologist)
P.S. If you know any gifted programmers, ask them if they can envision a program like this, and, if they say yes, if they’ll build it! It should be easy to get a Kickstarter (or other crowd-funded) project to kick this off. Now or in the future when it becomes more feasible.