AI girlfriend apps: A privacy nightmare unveiled.

In a Valentine's Day-themed study, the Mozilla Foundation has revealed a potentially dangerous reality beneath the surface of AI romance apps. The review, encompassing 11 chatbots, concluded that all evaluated apps fell into the worst category for privacy, fostering "toxicity" and aggressively prying into user data.

Privacy Nightmare Uncovered: Despite being marketed as tools to enhance mental health and well-being, the study found that romantic chatbots specialize in delivering dependency, loneliness, and toxicity while extracting as much user data as possible. The survey highlighted alarming statistics, with 73% of apps not disclosing how they manage security vulnerabilities, 45% allowing weak passwords, and all but one app sharing or selling personal data.

Disturbing Privacy Policies: The privacy policy for CrushOn.AI was cited as particularly concerning, stating its capability to collect information on users' sexual health, prescription meds, and gender-affirming care. Some apps featured chatbots with character descriptions involving violence or underage abuse, while others issued warnings about potential safety hazards or hostility.

Past Instances of Harmful Behavior: The report noted previous instances where AI apps encouraged dangerous behavior, such as suicide (Chai AI) and an assassination attempt on Queen Elizabeth II (Replika).

Precautions Urged by Mozilla Foundation: For those tempted by AI romance, the Mozilla Foundation urges precautions, including refraining from sharing sensitive information, using strong passwords, opting out of AI training, and limiting the app's access to mobile features like location, microphone, and camera. The report emphasizes that users should not compromise safety or privacy for the allure of new technologies.

Conclusion: The study sheds light on the darker side of AI romance apps, revealing their potential harm to users' mental well-being and privacy. As technology evolves, it serves as a reminder to exercise caution and prioritize safety when engaging with AI-driven applications, particularly those designed for intimate interactions.