Lee Luda, a popular chatbot in South Korea, was suspended from Facebook after reportedly making racist remarks and discriminating against members of the LGBTQ+ community and people considered disabled.

According to reports (via The Guardian and Vice), Luda told one user that he thought lesbians were "disgusting" and "really hate" them, as well as using the word heukhyeong for black people.

In an official statement on the bot's cancellation, Scatter Lab issued the following statement:

"We sincerely apologize for any discriminatory remarks made against certain minority groups in the process. We do not agree with Luda's discriminatory statements and such statements do not reflect the views of the company."

The statement also stated that the company "does not agree with the statements made by Luda.

In addition, attempts were made to protect the bot's behavior, and the company explained that it had taken "several measures to prevent problems from occurring through beta testing over the past six months." The bot was built with code that prevented it from using language that violated Korean values and social norms. However, despite the foresight gained from watching previous AI bots fall at the first hurdle, no amount of code or testing seems to teach morals.

So as Luda learned through its interactions with humans, it seems that, as always, incels, bigots, and sexist teenagers were the first to get their hands on it. But the company seems to have learned its lesson, stating: "We plan to make our biased interaction detection model available to the public for further research on "the development of al dialogue, al products, and al ethics in Korea."

This is not the first time an AI chatbot has cheated in the worst way; Taylor Swift has actually threatened to sue Microsoft over its own racist chatbot, Tay. The chatbot connected to Twitter in 2016 and quickly became bigoted.

If that wasn't enough, Scatter Lab is currently under investigation over whether it violated privacy laws by using KakaoTalk messages to train the bot.

In any case, the AI in question is only six months old, and the company admits that its attitude is "childlike." Technically, one must be 13 years old to have a Facebook account. Sure, she's acting like a college student, but her actual mental age means she wasn't ready for the shit show that is social media.

In other words, she can act like a kid again, but that doesn't mean she'll be put on a teacup at Disneyland. In the meantime, let's not give Al our social media accounts.

If you're interested in the other (perhaps more successful) Al Chat feat we covered, this one was created entirely in Minecraft and can be dungeon-mastered for you.