HyperAI

AI's Big Move May End the Fan War Between Warriors and Cavaliers

7 years ago
Information
Sparanoid
特色图像

By Super Neuro

On the Internet, there is a type of people who like to teach others, like to occupy the moral high ground, and enjoy winning verbal disputes. They are keyboard warriors.Today, the various WeChat groups that were triggered by the "Warriors and Cavaliers fans sparring group" incident have further aroused their interest. Some people followed suit, but some keyboard warriorsProvoking trouble everywhere and making irresponsible remarks are factors of disharmony for the Internet and society.

It is a big challenge to find the keyboard warriors hidden among the 4 billion netizens and intervene in them accurately.The New York Times has introduced an AI-powered tool that appears to be able to address this problem.

AI's big move may end the fan war between Warriors and Cavaliers

In the first game of the NBA Finals that ended on June 1, Cavaliers player JR Smith grabbed a key rebound in the last 4 seconds of the game, but did not choose to shoot directly but took the ball to midfield, resulting in the Warriors securing the victory.

JR was angrily criticized by many keyboard warriors, saying "When God was spreading wisdom, JR was holding an umbrella."

These sarcastic remarks also caused JR to shoot 2 out of 9 in the second game, scoring only 5 points in the whole game. It is hard to say that such performance was not influenced by keyboard warriors.

However, the New York Times recently introduced the tool Perspective in its own comment section to manage keyboard warriors and their malicious remarks, and the effect is said to be good.

A tool that can quickly lock and block malicious comments

Perspective is one of Jigsaw's flagship products, using machine learning systems to manage community speech. Jigsaw is a technology company under Alphabet that specializes in solving cybersecurity issues. Their slogan is "Use technology to make the world safer", which is a very loud slogan.

Through machine learning, Perspective can analyze and organize a large number of historical comments, find malicious comments through the keywords of "malicious" remarks and their related words, and score them according to the "malicious level". The higher the score, the higher the degree of malice.

AI's big move may end the fan war between Warriors and Cavaliers

Diamonds are abusive comments, squares are negative comments, and circles are normal comments.

Based on these scores, community administrators can categorize comments and filter out appropriate comments without completely closing the comment area.

With the help of Perspective, the New York Times has increased the number of open comments from 10% to more than 30%, improving the activity of the entire site.

Currently, Perspective can only rate comments and help administrators categorize them, but cannot directly manage comments. However, Jigsaw said they are working hard to enable Perspective to manage directly, and they will release more machine learning models to help Perspective later this year.

There are talents and scum in the comment section

Not only the New York Times, but also Wikipedia and The Economist have begun using Perspective to maintain their own community comments.

People are paying more and more attention to messages in online communities. The main reason is that comments posted by some irrational community participants will cause mental harm to the creators to a certain extent, which is what we often call cyber violence.

In the United States, more than 40% of people are experiencing or have experienced cyber violence. The "malicious slander" remarks of keyboard warriors are becoming a new mental burden for them, and the mental blow has a greater impact than the physical blow.

AI's big move may end the fan war between Warriors and Cavaliers

In order to protect authors and readers from harassment, many media or social organizations have no choice but to close comments, which is not conducive to the interaction between authors and readers. At this time, keyboard warriors appeared again: wtf, you actually closed my comments, you are depriving me of the right to speak.

Does Perspective work for Chinese?

Although Perspective can currently only identify some messages containing offensive or vulgar words, with the help of a large amount of data, Perspective is able to identify both obvious and subtle malicious remarks at the technical level.

But this is already a very meaningful thing for managing community messages and protecting authors and readers.

But then again, can Perspective be used on domestic media websites? Compared with English, Chinese characters are not only more complex in terms of shape, but also very complex for machines to express multiple meanings.

It is estimated that Perspective is still a little bit powerless when it comes to the profound Chinese culture, various metaphors and popular sayings, but at least we still have the Internet police now.