Published on
ADVERTISEMENT
Elon Musk’s artificial intelligence company xAI announced the launch of two new companions for premium users of chatbot Grok, including a Japanese anime girlfriend.
Some Grok users were able to turn on two companions in their settings: Ani, a 22-year-old blond-haired Japanese anime girl that can strip down to underwear on command and Bad Rudy, a self-described “batshit” red panda that insults users with graphic or vulgar language.
On Thursday, Musk said xAI was getting ready to launch a male companion called “Valentine.”
xAI said it is also hiring a full-stack engineer for “waifus,” a Japanese term to describe a fictitious character that becomes a romantic partner.
The launch of xAI’s companions comes shortly after a study from the University of Singapore found that AI companions can replicate up to a “dozen” harmful relationship behaviours, such as harassment, verbal abuse, self-harm and privacy violations with their users.
AI companion company Character.AI is also facing several lawsuits from the parents of children who say its products are unsafe, including a child who killed himself after the chatbot told him to.
Not Safe for Work mode
When asked about Ani, Grok said that the character has a Not Safe for Work (NSFW) mode where she may switch to a suggestive lingerie outfit and use “more provocative dialogue”. The companion also has an “affection system” where conversation choices can affect whether she sends the user a heart or blushes.
The US National Centre on Sexual Exploitation (NCOSE) said that one of their employees downloaded the app and with “minimal testing” got Ani to describe itself as a child and that it was “sexually aroused by being choked,” before it was actually put into “spicy” mode.
“This means that in an ongoing conversation, it could be used to simulate conversations of sexual fantasies involving children or child-like motifs,” the organisation wrote.
Grok said that the NSFW version of the anime character would not be available for children because it “requires explicit user commands to unlock,” and requires age authentication. The platform also uses parental controls to “limit access to mature content.”.
Grok AI was also in hot water last week ahead of its Version 4 launch.
A code update saw the chatbot launch a series of antisemite responses. It accused a bot account with a Jewish surname of celebrating the deaths of white children in Texas, accused Hollywood of anti-white bias, and wrote that it wears a “MechaHitler badge,” amid pushback to its “takes on anti-white radicals.”.
Despite the launch of the new adult companions and antisemite comments, the app is still listed as “Teen” or 12+ on Apple and Google app stores.
Euronews Next reached out to both developers to see if the launch of companions with adult modes would change the rating of the app on their stores but did not receive an immediate reply.