Blondes, with blue or green eyes, long hair and perfectly proportioned bodies are some of the virtual girlfriends that can be found in the numerous applications that can be installed on a smartphone with a simple click. But behind these apps lurk a multitude of dangers of which the user is often unaware.
According to the Mozilla Foundation, which analyzed a dozen of these romantic chatbots, 90% of these applications can sell their users' data or pass it on for advertising purposes. Some 73% do not disclose how they deal with security vulnerabilities, and almost half (45%) allow passwords as weak as simply entering the number 1. But the dangers do not stop there: Likewise, half of the apps (54%) do not allow the deletion of the personal data provided, which may include conversations with a virtual girlfriend. In fact, most of these apps make it clear that they can share the information given to the virtual partner with the government or law enforcement agencies without the need for a court order.
Major dangers
"Artificial intelligence brides are a nightmare for people's privacy and it's a bad idea to use them," clarifies Josep Curto, professor at the Faculty of Computer Science, Multimedia and Telecommunications at the Universitat Oberta de Catalunya (UOC), summarizing four of the biggest dangers in line with the Mozilla Foundation's warnings."These generative AI chatbots are designed to ask for as much information as possible, from photos to your voice to the most intimate information, such as sexual innuendo," he explains, using phrases such as "I want to know everything about you".
"This data is likely to become the main source of revenue for these companies, which share it with companies interested in profiling," says the UOC professor. He explains: "A third reason is that some of these applications often include trackers that send information to countries that have an interest in spying on people and who believe that privacy is not a right. And finally, there is no transparency about the nature of the underlying artificial intelligence model used in the system, nor whether it has been developed according to the premises of responsible AI."
Given this lack of transparency and the fact that the app market is doing "little" to limit the spread of these apps, the UOC professor calls for citizens to be made aware of the problems these virtual girlfriends can cause. "We need to avoid these apps so as not to get ourselves into complicated situations in the future, not only because of the data issue, but also because of possible fraud scenarios [a common scam on the internet]," Curto clarifies.
Psychological problems
But these romantic chatbots not only put the user's privacy at risk, they can also cause psychological problems. "The very fact that you mix AI with the word girlfriend - because basically most apps offer a female profile - seems dangerous to me," notes Sylvie Pérez, a professor at the UOC's Faculty of Psychology and Educational Sciences.
"The fact that she is a woman helps to perpetuate gender stereotypes of complacent, submissive women who are always available and never in a bad mood," she adds. Moreover, the UOC professor says, a girlfriend is created by an artificial intelligence based on the user's character. It is therefore a way of interacting with someone you imagine on the other side and who is ready to always give you a reason "by creating patterns of understanding relationships with others that do not correspond to reality": "You can confuse what social interactions are and what their real nature is, which is based on the fact that there can be conflicts, disagreements, different points of view... And that's what makes us intelligent, what ultimately makes us grow," the UOC professor recalls.
"There is also the risk that if you spend many hours interacting with these virtual friends, you will have difficulty distinguishing reality from fiction," she adds. "Either you understand from the outset that you are interacting with yourself, or you can lose your mind," concludes the psychologist.
