【】

2024-11-22 01:00:39 [知識] 来源：有聲有色網

When we spend so much of our time online, we’re bound to learn something while clicking and scrolling. Discover something new with Mashable’s series I learned it on the internet.

When I was a teenager in the 1990s, I borrowed a library book on the prophecies of Nostradamus. For weeks I annoyed my family with all the supposed evidence for its main claim, that the 16th century French mystic had predicted the end of the world in 1999. Luckily, the World Wide Web wasn’t really a thing yet, so after I returned the book my Nostradamus fever broke. Only years later did I look back and see my confirmation bias. I’d felt the now-familiar thrill of doing research, without realizing that what I was researching was bullshit — from mistranslated start to misinterpreted finish.

I thought about my Nostradamus mistake a lot during 2020, when the QAnon conspiracy theory came closer than ever to the mainstream. “We’re the only cult that tells you to do your own research!” Q fans would say when shrugging off accusations of cult-like behavior. Indeed, there was plenty of knotty research-like behavior to be found around every Q drop: tons of numerology, layer upon layer of commentary; a long line of leading questions. For adherents, it felt like putting together a jigsaw puzzle — so much so that they failed to see the picture they were building was light years away from reality.

The internet is the greatest tool ever created for learning, but also the greatest vector of misinformation. It offers a wealth of truth and an embarrassment of nonsense. We are susceptible to confirmation bias even if we’re not QAnon types. We’ve all been taken in by social media hoaxes more often than we care to admit, or skimmed over the surface of a subject and come away with the wrong impression.

Prime Day deals you can shop right now

Products available for purchase here through affiliate links are selected by our merchandising team. If you buy something through links on our site, Mashable may earn an affiliate commission.

iRobot Roomba Essential Vac Q011 Robot Vacuum Cleaner—$159.99(List Price $249.99)
Samsung Galaxy Tab A9+ 10.9" 64GB Wi-Fi Tablet—$142.49(List Price $219.99)
Apple AirPods Pro 2nd Gen With MagSafe USB-C Charging Case—$168.99(List Price $249.00)
Fitbit Charge 6 Fitness Tracker With 6-Months Membership—$99.95(List Price $159.95)
Apple Watch Series 9 (GPS, 41mm, Midnight, S/M, Sports Band)—$279.99(List Price $399.00)

“A little learning is a dangerous thing,” Alexander Pope famously wrote in 1709. Less remembered is Pope’s reason: When it comes to knowledge, “shallow draughts intoxicate the brain.” We might add so does drinking deeply from poisoned wells.

If you’d like to learn more on any topic without getting drunk, or just brush up your research skills before annoying your family again, here’s my advice, drawn from years in the internet trenches in the ongoing war between fact and falsehood. You shouldn’t just take my word for it, of course, so I’ve included links to further reading after each guideline.

And in case this is all too much to read, I've put the TL;DR in the headline: Never, ever trust a single source.

1. Don't just Google it.

In January 2014, Google introduced “featured snippets” — those little boxes at the top of search results that claim to answer your query without you even clicking on a result. But most snippets relied on an automated algorithmic process with little in the way of human curation or fact-checking. Websites got to work optimizing their articles so they’d show up in featured snippets. In 2017, one history professor found snippets claiming incorrectly that multiple presidents have been KKK members, and that President Obama was planning a coup.

Google has since cracked down on snippet misinformation, but it’s still far from an academically rigorous process. I’m not the only one to have found snippets with terrible medical advice. SEO experts are still gaming the system, for snippets and for search results as a whole. So are politicians: Witness Boris Johnson’s bizarre 2019 claim that he liked to paint cardboard buses as a hobby, which briefly pushed an infamous lie-covered bus he rode during Britain’s EU referendum out of the top Google results for “Boris Johnson bus.”

Google can still be a useful research tool — Google Scholar especially — but it’s no magic mirror that automatically answers your questions. (Nor, by extension, are Google Home, Alexa, or Siri, all of which rely on Google.) You’d be better off ignoring snippets altogether, and treating the ranking of search results with a healthy dose of skepticism. Also, here’s a wild idea: Maybe take a look at the second, third or even fourth results page before deciding which links to open.

Read more about the most egregious Google snippets, why three searches over time show that Google results are getting worse, and how to reduce your reliance on algorithms.

2. Wikipedia is pretty good, actually

The crowdsourced online encyclopedia gets a bad rap, thanks to the edit wars that raged in its early days, when any internet denizen could come along and add anything with little scrutiny. But at the age of 20, Wikipedia has matured into a mostly trustworthy source with fairly rigorous editing.

Mashable Light SpeedWant more out-of-this world tech, space and science stories?Sign up for Mashable's weekly Light Speed newsletter.By signing up you agree to our Terms of Use and Privacy Policy.Thanks for signing up!

Try to vandalize a popular entry today and you’ll either find yourself locked out, or your change quickly switched back by hawk-like volunteer editors. Pages on climate science are particularly well-defended. Miscreants aren't entirely anonymous; IP addresses reveal when Members of Congress, for example, appear to be editing their own entries. Fears of automated scripts that would game the system, SEO style, have failed to materialize; in fact, bots seem to be better at defending Wikipedia than rewriting it.

Studies of its accuracy are mixed, as the “reliability of Wikipedia” page notes in comfortingly comprehensive detail. (One thing that helps increase trust in any given online source: its ability to openly and accurately criticize itself.) But most of the negative studies hail from the site’s first 10 years of life. In the past decade, researchers have found accuracy similar to offline textbooks. For example, a widely-cited 2014 study found that Wikipedia entries on drugs were 99.7 percent accurate when compared with a pharmacology textbook.

Where errors exist, most seem to be errors of omission. Wikipedia editors are still too male and too white to accurately encompass the totality of human knowledge and cultural history. (Some women and people of color are working hard to fix that.) Still, the risk of running across misleading entries is real. You absolutely shouldn’t use Wikipedia as a primary source for term papers, essays, articles or anything else official; at Mashable, we rarely allow Wikipedia links in stories.

But as a way to get your head around the basics of a topic, or to find a preliminary answer to casual questions that come up at the dinner table, you could do a lot worse. In today’s storm-tossed sea of crazy, the Wikipedia page is a functioning first port of call. Just be sure to click through on the citations — many book-based citations now actually lead to the book, thanks to the Internet Archive — and look askance at any statement that lacks a source.

Read more on Wikipedia's war on climate denial, the bot-on-bot edit wars, and a conspiracy theory that survived on Wikipedia for a record 15 years.

3. YouTube is a trash fire.

At the other end of the spectrum from Wikipedia is the Google-owned video service where users are literally given cash to produce controversial content. The radicalization of YouTube is a well-documented phenomenon: Its algorithm rewards time spent on any video, regardless of its accuracy, with higher placement in the "up next" bar. So if you want that sweet, sweet ad money, it doesn’t pay to be dull and risk your fans clicking elsewhere; rather, it pays to scream outrageous lies at the top of your lungs.

YouTube CEO Susan Wojcicki isn’t the only tech leader to let misinformation run rampant, but her enforcement does seem to be the most lax. For example, while Trump was banned from Twitter and Facebook in the wake of the Capitol attack, leading to a massive drop in disinformation, YouTube merely prevented him from posting new videos. His 46-minute screed filled with baseless accusations of a stolen election is still available. Likewise, even if YouTube did belatedly take down “Plandemic,” it still took a year for the service to actively fight back against COVID-19 misinformation.

Sure, there are still plenty of legit academic sources on YouTube that can help accelerate your learning. Udemy, Khan Academy, and Crash Course are some examples of generally trustworthy education channels. The quality of TED and TEDx talks vary wildly; in general, the TED format puts too much focus on one point of view, and would be a lot better if it had a devil's advocate. Your mileage may vary, depending on the subject and the likelihood that someone might want to lie to you about that subject.

Personally, I wouldn't even believe a video on how to tie your shoelaces without getting a second source. That video your uncle sent you on some political topic or other? Forget about it.

Read more on the making of a YouTube radical, the kid who cracked the YouTube algorithm's secrets, and how the service swung the election in Brazil to an autocrat.

4. The media is a mixed bag.

The proliferation of media is probably the greatest cause of confirmation bias in the online world. With so many articles from so many sources appearing in search results, and no way to sift between them other than their search ranking, it makes sense that the average reader would just click on a headline that aligns with their world view.

Blindly believing legacy media isn't an option either. Anyone who remembers Stephen Glass or Jayson Blair or Judith Miller isn't going to automatically trust the New Republicor New York Times, respectively. A random no-name blog could be semi-literate nonsense, or a giant bag of lies, or a well-sourced journalist who just doesn't want to work for the big guys (the now famous "pivot to Substack"). Increasingly, it's a crapshoot. Even those graphs that put news sources on a left-right spectrum aren't necessarily fact-checked.