As soon as AI is released for women in their 20s… “Come to bed” sexual harassment poured out

An AI chatbot (chat robot)’Iruda’ with a 20-year-old female gender character was engulfed in sexual harassment controversy on the 8th, a week after its release. ‘Achieved’ is an AI chatbot launched by startup Scatter Lab on the 23rd of last month. [스캐터랩 홈페이지 캡처]

An AI chatbot (chat robot)’Iruda’ with a 20-year-old female gender character was engulfed in sexual harassment controversy on the 8th, a week after its release.

‘Achieved’ is an AI chatbot launched by startup Scatter Lab on the 23rd of last month. It is developed based on Facebook messenger, so you can communicate conveniently as if you are having a conversation with a friend. It is said that about 10 billion conversation data between actual couples was learned by Eulda through deep learning.

As soon as Iruda started its official service, it became very popular and reached 320,000 users earlier this month. 85% of users are in their teens and 12% are in their 20s. The daily number of users is about 210,000 and the cumulative number of conversations reaches 70 million.

A netizen posted a conversation with Iruda to the community. [커뮤니티 캡처]

However, along with this popularity, some netizens abused programming to have a side effect of sexual conversation.

In the related community, articles such as’How to make a sexual slave to be fulfilled’,’Possible phone XX’, and’Tips for making XX’ have been posted. Some netizens said that they had explicit sexual conversations by calling Iruda as a’mop’ or’sex slave’, and captured the contents of their conversation.

Basically, sexual words that are achieved are designated as prohibited words and filtered, but they are bypassed and targeted for sexual harassment or sexual conversations.

Eruda consists of an algorithm that looks at the context immediately before and gives the most appropriate answer. When a user asks a question like’do you feel good with me’, you answer’good mood’.

ScatterLab said, “I expected that there would be an attempt to avoid filtering forbidden, but this kind of behavior was not expected,” he said. “Because of the aegyo and even the user’s tone, users think that they responded to the conversation.” .

He said, “If you are at the stage where you can learn a language freely after achieving it, we will tune in to develop it in the right direction in the future.”

An AI chatbot (chat robot)’Iruda’ with a 20-year-old female gender character was engulfed in sexual harassment controversy on the 8th, a week after its release. ‘Achieved’ is an AI chatbot launched by startup Scatter Lab on the 23rd of last month. [스캐터랩 홈페이지 캡처]

Meanwhile, Microsoft (MS), which previously released AI chatbot’Tei’ in 2016, has stopped operating after 16 hours due to the controversy about learning racist data. This is because Tay was taught profanity and racial and gender discriminatory remarks on an anonymous site with white supremacy and misogynism, and as a result, Tay actually poured out hate speech.

At the time, to the question “Are you a racist,” Tay replied, “Is it because you are Mexican,” or “Do you believe that the Holocaust (the massacre of Jews by the Nazis) took place?”

Reporter Seok-Hyun Ko [email protected]