Contact
Subscribe
Advertise
Follow us on:
Search:
Banner
Fake News How AI can Help Debunk it!

Fake News
How AI can Help Debunk it!

By Stella Law

TP 201904 future 02鎻浜哄伐鏅鸿兘鏄浣曞彂鐜板亣鏂伴椈鐨

1475骞, 鍦ㄤ竴涓編濂界殑鏃╂櫒锛屽湪鎰忓ぇ鍒╃箒蹇欑殑鐗逛鸡鐗归晣鏈変釜璋h█锛屽綋鍦扮姽澶ぞ鍖虹殑棰嗗浜鸿皨鏉浜嗕竴鍚嶅悕鍙タ钂欒鐨勫熀鐫e緬濠村効銆傝埃瑷琚涓哄悎娉曟柊闂伙紝骞跺鑷磋甯傚嚑鍚嶇姽澶汉閬埌閫崟鍜屾姌纾ㄣ


1835骞8鏈堬紝绾界害鈥滃お闃虫棩鎶モ濆彂琛ㄤ簡涓绯诲垪鐨勫叚绡囨枃绔狅紝鎻忚堪浜嗘湀鐞冧笂鐨勭敓鍛藉拰鏂囨槑銆 杩欎簺鏁呬簨褰掑姛浜庡綋鏃朵紵澶х殑澶╂枃瀛﹀绾︾堪璧瓏灏旂埖澹傝繖浠朵簨鍦ㄥ巻鍙蹭笂琚О涓哄ぇ鏈堜寒楠楀眬銆

 

涓婇潰鐨勪袱涓緥瀛愯〃鏄庡亣鏂伴椈鍦ㄦ垜浠殑绀句細涓竴鐩存槸涓涓棶棰樸

 

浜哄伐鏅鸿兘鐩墠姝e浜庢墦鍑诲亣鏂伴椈鐨勪竴绯诲垪鎺柦鐨勬渶鍓嶆部銆傚埄鐢ˋI璇嗗埆铏氬亣鏂伴椈鐨勮兘鍔涗篃鏈夊姪浜庤瘑鍒櫄鍋囧浼犵殑妯″紡鍜屼俊鎭

 

鍦ㄤ汉宸ユ櫤鑳藉拰澶ф暟鎹伐鍏峰彉寰楁洿鍔犲鏉傚拰鍙潬涔嬪墠锛屾垜浠渶瑕佷笓娉ㄤ簬鏁欒偛鑷繁鎴愪负鎵瑰垽鎬ф濊冭咃紝鎴栬呰嚦灏戜笉瀹规槗涓婂綋閭d簺铏氬亣淇℃伅鍜屽浼犮傚繀椤婚槻姝㈢浉淇′换浣曡镐汉鍚椈鐨勬柊闂汇

TP 201904 future 05The year was 1475. A rumour broke out in the busy town of Trent, Italy, on one fine morning, that the leader of the local Jewish community had murdered a Christian infant, named Simonino. The rumour was taken as legit news on the face value and it resulted in the arrest and torture of several Jews in the city. Pope Sixtus IV himself tried to intervene, but the story had spread like wildfire.

 

In August 1835, a series of six articles were published in The Sun, a New York newspaper, citing the discovery of life and civilization on the Moon. These stories were attributed to Sir John Herschel, a great astronomer of that time. This incident went down in history as the Great Moon Hoax.

 

The two examples above show that fake news has been historically a problem in our society. If we were to guess, it has been probably around for as long as our early ancestors learned the fine art of gossiping and rumour mongering.

 

Fake news is far worse a social menace today than ever, because of the easy access to (often unverified) information facilitated by the digital sphere -- especially social media. Sure, the government can always go for a crackdown on social media and fake news propagators, although that would perhaps count as an assault on free speech in much of the democratic world.

TP 201904 future 03So what do we do? Can鈥檛 we find a better way to tackle fake news?
Well, as it turns out, artificial intelligence (AI) can be of great help on that front.

 

Artificial Intelligence to the Rescue

AI is, at the forefront of a slew of measures, currently being considered to combat fake news. Harnessing AI鈥檚 capability to identify fake news can also help in recognizing patterns and information that are a part of false propaganda.

 

The idea is to let AI learn from data, identify patterns and reach its own conclusion with minimal human intervention.

TP 201904 future 04How?

Well, for example, a joint project by MIT鈥檚 CSAIL (Computer Science and Artificial Intelligence Lab) and the QCRI (Qatar Computing Research Institute) was announced last year in October that claimed to identify sources of fake news, even before they can spread out.

 

This is a potential solution involving automatic classification of news sources that are unreliable and helping out fact checkers.

 

However, detecting fake news with or without AI is not an easy task. Websites, like Snopes and Politifact, are constantly working to expose false/fake news before they can inflict too much damage. However, if something leaks once, it is likely to spread around like a virus, before the systems in place are able to stop it.

 

The system aims to work from the base by looking at the website, Wikipedia page, URL structure, Twitter account, as well as its web traffic. It also looks for keywords and language that might be indicative of a strong political bias (in case of political news outlets).

 

Researchers first took data from Media Bias/Fact Check (MBFC), a website that has human fact checkers analysing the authenticity of over 2000 news websites.

 

Thereafter, it fed the data to a specific machine learning algorithm and programmed it to classify news items in a way similar to MBFC. The system was given a new news outlet and it performed with 65% accuracy in detecting the level of the factuality of the news website. Moreover, it was approximately 70% accurate in determining if the news was left-leaning, right-leaning or even moderate.

 

The team ascertained that probably the most reliable method of detecting fake news was to observe the common linguistic features across the news pieces, which include the complexity, structure, and sentiment.

 

For instance, websites giving out fake news are more likely to use language that is emotional, subjective and using expressions in exaggeration.

TP 201904 future 07AI Tools to Spot Fake News

In 2015, Google published a paper about a method of scoring websites on the basis of the accuracy of the facts presented. This algorithm assigns a specific trust score to a document that is used in the total scoring process by Google.

 

There are various other tools coming up that can help in fighting fake news. Below is a list of a few of them:

 

鈼 Spike: Spike is a tool for spotting, as well as predicting viral events and stories.
鈼 Google Trends: A Google service that watches trends by analysing top search queries.
鈼 CrowdTangle: A content discovery and social monitoring platform.
鈼 Le Monde鈥檚 Le D茅codex: It is a database of over 600 news websites that have been identified as 鈥榝ake鈥, 鈥榮atire鈥, 鈥榬eal鈥, etc鈥
鈼 Hoaxy: An online tool that helps people to check facts online.

TP 201904 future 06Apart from the above, various other tools are also available that can enable people to check and identify even the most outrageously fake news stories. Many times, the problem lies in users, who are generally disinclined to make an extra effort towards checking the authenticity of a story.

 

Conclusion

Until AI and Big Data tools become more sophisticated and reliable, we need to focus on educating ourselves to become critical thinkers, or at the least, less gullible to dubious information and propaganda. The willingness to believe any sensational piece of news must be prevented.

 

An MIT study from last year found that real news took 6 times longer to reach Twitter users and that fake news was likely to be retweeted 70% more than real news!

 

Fake news is an ever growing and compounding menace in our world. With AI being used to combat it, the future looks good for more sophisticated tools that help further the cause.

  About TJ+     Contact TJ+     Subscribe to TJ+     Advertise on TJ+  
Copyright © 2019 TianjinPlus.com All rights reserved.