The personal data of 106,362 users of the MyLovely.ai adult AI companion app has allegedly been compromised, with a large volume of explicit messages among the exposed data.
The alleged breach is the latest reminder that when you’re discussing intimate sexual fantasies and commands with your AI companion character, you shouldn’t expect your sexual prompts to stay private. A recent report highlighted how many AI companion apps weren’t designed with even the most basic security, and the alleged MyLovely.ai debacle hardly suggests that change is imminent.
The email addresses, user IDs, account creation dates, exact written prompts and potentially the social media handles of 106,362 MyLovely.ai users were allegedly compromised. Explicit images generated by the platform’s AI characters within private chats were also allegedly part of the breach, the alleged breach was posted as a notification on a cybercrime forum.
Around 70,000 explicit prompts written by users, linkable to specific user IDs, were allegedly part of the leaked data. This raises the prospect of the data being used by cybercriminals, if the usernames can be linked to real identities.
Malwarebytes, the security firm that flagged similar risks after the BangBros leak, said MyLovely.ai users should expect doxxing and sextortion attempts.
“We are aware of the reports. Nothing critical or highly sensitive was leaked — no passwords, no payment data, nothing like that,” a MyLovely.ai spokesperson told SEXTECHGUIDE. The company did not respond to follow-up questions asking for confirmation of exactly what data had been exposed.
The response somewhat sidesteps the fact that most people would consider email addresses as personally identifiable information. User IDs tied to explicit sexual prompts are personally identifiable information, particularly when combined with social media handles. Malwarebytes’ own assessment is that users should brace for doxxing and sextortion attempts, which is not what ‘nothing critical’ sounds like.
MyLovely.ai lets you interact with the platform’s stock female AI characters, in sexual scenarios that push well past what most platforms would permit. In one scenario you simulate locking a 19-year-old woman in a room with you, against her will, then masturbating as she gets angry and wields a baseball bat.
The platform hosts extensive AI-generated porn, showing both photo-realistic and anime characters completely nude and in explicit sexual scenarios. You can also create original AI characters from scratch and share them on the platform.
The particularly explicit style of MyLovely.ai suggests that there’ll be some pretty out-there private messages in the data breach. In 2024 Muah.ai, another hardcore AI companion chatbot, suffered a user data breach that included users’ private sexual prompts to the chatbot.
It’s not just the the AI companion chatbots at the most explicit end of the spectrum that have been shown to have highly-questionable security, though.
The recent report by app security firm Oversecured showed that many of the most popular AI companion chatbots, that have collectively been downloaded around 150 million times, had flawed security making them at risk from hacking and data breaches.
For all the touting of top security measures, when dealing with any kind of AI chatbot the adage should remain: Don’t write or say anything on the internet that you wouldn’t want to be posted for the whole world to see.
Leave a Reply