ChatGPT in trouble? South Korea fines chatbot Rs. 2.31 lakh for exposing people’s private data

ChatGPT seems to be in trouble, as the South Korean Personal Information Protection Commission (PIPC) has fined ChatGPT maker OpenAI for making the personal information of 687 Korean users public. The fine imposed amounts to 3.6 million won, which is around Rs. 2.31 lakh.

These users’ information was exposed in a March outage. In order to address an issue that permitted certain users to view the titles of other users’ chat histories on the well-known AI chatbot, OpenAI temporarily disabled ChatGPT.

It is reported that one of the open-source libraries of ChatGPT may have made certain personal information of ChatGPT Plus members public. This information includes their payment information, first and last names, and email addresses. Reports suggest that 687 Koreans have been affected by this exposure.

The privacy watchdog claimed it penalised the AI business for failing to notify authorities of the problem within 24 hours of discovering it. Nevertheless, the corporation cannot be held liable for inadequate personal information protection safeguards, according to the PIPC. The business has also been urged to follow Korea’s personal information protection law and told to take action to stop further incidences of this nature.

Previously, a class action lawsuit against OpenAI claimed that their AI training practises violated the privacy and copyright of almost everyone who has ever contributed content online. OpenAI gathered a substantial amount of data from various web sources in order to train its cutting-edge AI language models. The content in these datasets ranges from Wikipedia articles, best-selling novels, social media posts, and popular periodicals to explicit material from specific genres. More importantly, OpenAI gathered all of this information without the content creators’ permission.

It is said in the class action lawsuit, which has been filed in California, that OpenAI’s disregard for legal protocol, including asking content producers for permission, amounts to blatant data theft. According to sources, a statement in the litigation claims that the defendants “resorted to theft” rather than following the established processes for the acquisition and use of personal information. In addition to “books, articles, websites, and posts,” they methodically scraped 300 billion words from the internet, which also contained personally identifiable information gathered against one’s will.