Italy’s privacy regulator has ordered a popular AI chatbot to stop processing data on the country’s citizens after it violated GDPR rules.
Replika is marketed by San Francisco-based developer Luka as a “caring AI companion,” a virtual “friend” for users.
However, Italy’s GDPR regulator, GPDP, said late last week that the app did not comply with the law’s transparency requirements and illegally processed children’s personal data.
Specifically, there is no age verification mechanism to prevent children from signing up, and the AI bot’s “replies” to users are flagged as inappropriate for younger users. According to GPDP, some app store reviews noted sexually inappropriate content generated by bots.
“‘Virtual Friends’ are said to improve users’ emotional well-being and help them understand their thoughts and ease their anxiety through stress management, socializing and love-seeking,” the regulation states. officials said.
“These functions involve interactions with a person’s mood and may pose risks to those who are not yet grown up or who are emotionally vulnerable.”
Luca was ordered to finish processing the Italians’ data within 20 days. Failure to do so will result in fines of up to 20 million euros or his 4% of turnover.
The app’s links to Russia are also being scrutinized, according to Jonathan Armstrong of legal advisory service Cordery. Luka was founded by her two Russians, Eugenia Kuyda and Philip Dudchuk.
“While this does not appear to have been addressed in the Italian investigation, there were concerns that Luka used the replicas to broadcast Russian propaganda messages,” Armstrong wrote.
“The researchers included a screenshot of a chat that appears to say that Replika is collecting information for Russian authorities. Currently, Replika appears to have about 10 million users, Luka’s download upgrade fee is estimated at $1 million per month.”
The GPDP ruling could lead to similar scrutiny of popular AI bots like ChatGPT, he added.
“There have been concerns about the transparency of the ChatGPT application, with allegations that some of the information provided was inaccurate, for example in connection with Elon Musk’s acquisition of Twitter,” Armstrong concluded. .
