Killing Fake News Dead on Taiwan’s Most Popular Messaging App
Exploding in popularity even before they finished programming it, the hackers from Taiwan’s g0v community created an internet bot that checks "fake news" on the messaging app LINE.
Original reporting by Chih-hsin Liu
With translation and additional files by Aaron Wytze
Did you know that the Taiwan government has ordered the police to open fire on elderly protesters demonstrating against planned pension reforms?
How about the Taiwan government taking control of a privately-held agriculture board so it can import radioactive fruits and vegetables from Japan’s Fukushima region?
Or the news that Taiwan President Tsai Ing-wen nearly fathered the illegitimate baby of ex-president Lee Teng-hui?
These are all “fake news” stories circulated on LINE, the most popular instant messaging app in Taiwan.
As Facebook struggles to control the spread of incredulous news stories in the United States, LINE in Taiwan is facing a similar barrage of misleading or factually incorrect news items that seek to influence the public.
US media publications like Slate and New York Magazine have created web browser add-ons that alert you when you see a news story that is potentially fake while scrolling through Facebook, but LINE has been slow to make similar fake news alerts of their own.
That is, until now. Taiwan’s largest civic tech community “g0v” are creating a bot that helps LINE users determine whether the news item they viewed is intentionally misleading, or false.
The bot is called Cofacts, or zhende jiade in Standard Mandarin — an expression that denotes disbelief or shock.
Noticing a recent rash of false news stories shared on LINE that were designed to persuade old people in Taiwan to act in a specific way, or form a certain opinion, the makers of the bot tried to find a simple way for LINE users to check the veracity of news they are suspicious of.
Although the LINE bot is still in the very early stages of development, it has already attracted substantial attention from the Taiwanese public and local media. As the Taiwan public has discovered over the last year, the “post-truth” era is not only limited to the US.
“Fake News” in Taiwan, and “Fake Lettuce” from South Korea
LINE is an easy to use instant messaging app with a slick interface. Released in 2011 by South Korea’s Naver Corporation, the company claims to have over 700 million users, mostly in Asia. Taiwan was an early adopter of LINE, with an estimated 17 million Taiwanese people using the app by 2013 (about 75% of Taiwan’s total population).
It’s since been adapted by the young and old alike, and has become the primary mode of communication for many families in Taiwan.
Like most messaging apps, LINE allows users to form group chats with up to 200 people. Although LINE has the ability to bridge generational gaps and bring people together, its group chat feature can close off discussion too.
In Taiwan, where society can sometimes see heavy stratification along political party lines and cultural identity, LINE groups can become hotbeds of intense political discussion, rumors and gossip.
These chat groups are ideal places for fake news items to fester, in part because older generations of Taiwanese are perhaps not adept at double-checking misleading or false claims through web search engines like Google or Yahoo!.
Because of LINE’s high penetration rate in Taiwan, a fake news item can spread quickly, and shape perspectives or reinforce preconceptions.
For example, one of the most widely shared news items by Taiwan LINE users is a chain message showing a man creating “fake vegetables” in a Youtube video. The chain message reads in Chinese characters:
“After you see this, do you dare go to South Korea? Use South Korean products? All the food in South Korean restaurants is fake! Disgusting! It will certainly kill all you Taiwanese. I hope Taiwan doesn’t go through another poison food scandal like this, we’ve already had enough!”
Clicking the link takes the viewer to a Korean Youtube video of a man using an unknown material to make life-like food items, like a head of lettuce, and a fried shrimp. The video has already been viewed more than 3.9 million times.
In truth, the chain message and the video have nothing in common. The video was shot by South Korean television station SBS, and the man making the fake vegetables is a Japanese artisan who specializes in making inedible plastic food displays for local restaurants.
Becoming famous before the programming was finished
As incidents like the one above become more frequent, the desire for a news “fact-checker” has grown among Taiwanese netizens.
Soon after, the makers of Cofacts placed a request on g0v’s Facebook fan page for more collaborators to help out with programming, the post attracted over 3,000 “likes” and was shared nearly 2,000 times. Taiwan’s United Daily News and the China Times newspapers jumped on the story, writing hurried copy about its development.
“Just please stress in the article, that we are still in the process of working with programmers and editors to write code for the bot,” quipped Johnson Liang, a g0v member, and a Cofacts co-creator.
Liang has specifically created the LINE bot for older generations of Taiwanese who are not adept at using search engines like Google to check the veracity of suspicious news items or links.
The LINE bot is a simple internet bot that can be added within the LINE messaging app by scanning the following QR code. Once you have the bot in your chat log, you can copy and paste a suspicious website link into the LINE bot chat window. You can also directly ask the bot a question about a recent news event that has drawn skepticism. The bot will automatically send a response back if they believe that the news item is suspected of being inaccurate or misleading, or respond to your question with a relevant fact-checked response.
Like most internet bots, the LINE bot runs automated tasks, that are structurally repetitive. For the bot to work correctly, it needs a large amount of data to return back the correct response. As the number of fake news items accumulates, it also needs a huge pool of editors to continually make new entries for the bot.
Liang had recruited over 30 volunteer editors, and has since begun work on verifying the over 6,000 news items and links sent to the team. He hopes to empower Taiwanese elderly people to question misleading news items they receive in LINE group chats, and give them accurate responses to their queries.
But so far, only 46 percent of the response given by the LINE bot are accurate. For the team to improve the accuracy of the bot’s automatic responses, they need to constantly keep pace with new “fake news” items, as they arise. “This project will only be possible if we challenge ourselves to make the search function better,” said Liang.
Liang expects that it will take another year to complete the first phase of the Cofacts bot. In the future, Liang hopes he can add more functions that would give users the opportunity to exchange points of view and discuss their core values.
The LINE bot team have set a monumental task for themselves. How likely will the project come to full fruition? “That all depends on how much everyone really needs this tool,” says Liang.
If you’re interested in receiving updates about the Cofacts LINE bot, check out their public group on Facebook.
Copy Editor: Mini Wu
This piece can be used under the following copyright terms:
Within the first 48 hours of posting, this article is released under the CC BY-NC-ND 3.0 Taiwan license.
After 48 hours, this article is released under the CC BY 3.0 Taiwan license.
