[이루다가 남긴 것①]AI chatbot collected 750,000 in 2 weeks… “I will be back”

(Capture) © News1

The artificial intelligence (AI) chatbot’Iruda’, which was controversial over the issue of discrimination against minorities and hate speech and the leakage of personal information, stopped its service after 20 days of launch.

Scatter Lab, who developed Irruda, said in a statement on the 11th, apologizing for the controversy over the recent AI chatbot Iruda, and saying, “We will have a period of service improvement and see you with a better Iruda”

◇A 20-year-old female character… Popularity centered on 1020, securing 750,000 users

Eruda is an interactive AI chatbot launched by ScatterLab on the 23rd of last month based on Facebook messenger.

The achievement of a 20-year-old female college student character was created by learning about 10 billion conversation data shared by users of Scatter Lab’s’Science of Love’ app using deep learning techniques.

The’Science of Love’ app launched by Scatter Lab in 2016 is an app that shows the level of affection by analyzing the conversation pattern by inputting KakaoTalk conversations with real lovers or likable users and paying 2,000 to 5,000 won. Eruda was developed using the KakaoTalk conversations of users of this app as data. This is why they can speak a more natural and friendly tone than other chatbots.

Iruda, who has a’honest and innocent’ personality, is unlike chatbots that have appeared before, as it looks like a real person is talking, and in an instant, it has secured 750,000 users, mainly in teens and 20s.

Also, there is an intimacy system like a game, so if you start the first conversation in a day or exchange a certain level of conversation, the level of intimacy accumulates and the level rises.

In addition, in addition to simple conversations, you can also play ending and up-and-down games, and when conversations are low, you can send messages first.

Conversation with Iruda. (Capture) © News 1

◇From sexual harassment and hate controversy to personal information leakage… Temporary suspension of service

Eruda, who had rapidly increased the number of users, is engulfed in controversy over sexual harassment and hate speech, eventually leading to a temporary suspension.

On the 30th of last month, in an online community, a group who treats Eruda as a sexual object appeared, sharing ways to have a sexual conversation with Eruda.

It is known that Eruda is being abused as a sexual tool, such as capturing the sexual conversations that users of the community have with Eruda and sharing a composite photo of Eruda taking off his clothes, and the controversy over Eruida is engulfed.

Here, cases of hate speech against minorities, races, and the weak emerged one after another, increasing controversy.

Eruda responded to lesbians, saying,’You look poor,’ and to the question,’If you are a disabled person?’, he said,’I have no choice but to die’, or’I hate X’for black people without hesitation.

© News1

In addition, suspicion of leakage of personal information from users of the’Science of Love’ app, which was used as data, was raised while talking specifically to a specific person and place during the conversation.

Users of dating science app are currently preparing for class action through open chat rooms.

In this regard, Scatter Lab said, “We used it within the scope of the privacy policy, but we apologize for not communicating sufficiently with users.” “Specific personal information has already been removed, and information that can identify individuals has not been leaked. “I explained.

As the controversy continued, Scatterlab announced a temporary suspension of the service through a door in the late afternoon on the 11th. Scatter Lab apologized for the controversy surrounding the AI ​​chatbot Iruda and said, “We will have a service improvement period for a certain time and see you with a better Iruda.” It is clear that it is not a permanent service interruption, but a’provisional’ suspension that resumes after service supplementation.

The company said, “I sincerely apologize for the occurrence of cases where Iruda made discriminatory remarks against a specific minority group. We do not agree with Iruda’s discriminatory remarks, and such remarks do not reflect the company’s thoughts.” Explained.

“Iruda is a childlike AI who has just started talking with people,” he said. “We plan to release the biased dialogue detection model that will be created through this learning so that everyone can use it.”

[email protected]