Iruda developer “Apologizes for information disclosure…Delete personal information for desired users”

Scatter Lab, the developer of the artificial intelligence (AI) chatbot’Iruda’, apologized for faithfully working on an investigation into the alleged personal information leakage. In the process of sharing development records related to achievements on the open source sharing platform’GitHub’, he admitted that he did not properly protect personal information.

Scatter Lab issued an apology in the late afternoon on the 13th, saying, “I wasn’t careful about data management. We sincerely apologize for the exposure of sensitive information.” He added, “We will establish a fact-finding committee to thoroughly investigate all controversial matters, while establishing a system to protect personal information, including external experts, and conducting reinforced security training to prevent recurrence.”

It was released on December 23, last year, set as a 20-year-old female character. It also recorded 750,000 users within two weeks of its launch. However, controversy over sexual harassment over the achievements arose in online communities.

In addition, after hate speech against social minorities such as homosexuality and the disabled, the data used while learning KakaoTalk conversations of romantic science users were not properly anonymized, and the service was temporarily suspended on the 12th, until suspicion of personal information leakage.

Scatter Lab said, “I hope that this issue will not affect many fellow companies, researchers, and partners in the AI ​​industry.” “I will try to meet the social consensus on AI ethics.”

Scatter Lab collected KakaoTalk data from users and used it for the development of Eruda with the love analysis app’Science of Love’, and in this process, it is suspected of violating the Personal Information Protection Act. In addition, the development record was shared on the shared platform’GitHub’, but it turned out that it was not properly anonymized.

ScatterLab said, “It was confirmed that the open source released on GitHub contained internal test samples. The real name was automated and de-identified, but there were some parts that were not filtered out in the filtering process. We sincerely apologize for the disclosure of sensitive information. The GitHub post was immediately closed. There is no room for estimation of the relationship or life span of the people we spoke to.”

Regarding the suspicion that the consent to use personal information was neglected, “I decided that there would be no legal problem by obtaining consent to the collection and use of personal information in the same way that domestic and foreign services actually adopt.” I apologize again and again.

Scatter Lab announced that if users do not want to be used for AI, they will take active measures to prevent them from being used in future DBs, including deleting the DB in accordance with the Personal Information Protection Act.

In addition, in the process of collecting users’ KakaoTalk conversations, regarding the testimony that employees shared in the in-house chat room, “As a result of the investigation by a fact-finding committee, it has not been found yet.” “I will take the necessary measures and hold the relevant people strictly accountable.”

Meanwhile, science users in love are preparing for a class action lawsuit, saying that they did not know that their information would be used to achieve. The Personal Information Protection Committee and Internet Promotion Committee were put into the field on the 13th and initiated an investigation. Scatter Lab said, “We are working on the investigation sincerely at the request of the relevant agency, and we will do our best in the future.”

Newsis>

[ⓒ 세계일보 & Segye.com, 무단전재 및 재배포 금지]