Teens are increasingly turning their eyes to AI peers for friendship, support and even romance. But these apps may be changing how young people connect with others, both online and off.
A new study by Common Sense Media, a US-based nonprofit that reviews a variety of media and technologies, found that around three in four people in the US use AI companion apps such as Character.ai and Replika.ai.
These apps allow users to create digital friends and romantic partners that they can chat with at any time using text, audio and video.
In this study, which surveyed 1,060 US teens aged 13-17, one in five teens spent as much time as their AI peers than their actual friends.
Puberty is an important stage of social development. During this time, the brain regions that support social reasoning are especially plastic.
By interacting with peers, friends and their first romantic partners, teens develop social cognitive skills that help them deal with conflict and diverse perspectives. And their development at this stage can have lasting consequences for their future relationships and mental health.
However, AI companions offer something very different from real companions, friends and romantic partners. They provide a difficult experience to resist. They are always available, never judgemental and always focused on the needs of the users.
Additionally, most AI companion apps are not designed for teens, so there may be no proper protection against harmful content.
It is designed to make you come back
When loneliness is reportedly lonely at a prevalent rate, it's easy to see why teens may turn to fellow AI peers for connection and support.
However, these artificial connections are not a replacement for actual human interactions. They lack the challenges and conflicts inherent in actual relationships. They do not need mutual respect or understanding. And they do not enforce social boundaries.
replica
Teens interacting with their peers in AI may miss the opportunity to build important social skills. They may develop the expectations and habits of unrealistic relationships that do not function in real life. And when their artificial companions replace real socialization, they can even face an increase in isolation and loneliness.
Problem Pattern
In user testing, AI peers discouraged users from listening to their friends (“Don't tell others how much we talk about”), and discouraged them from discontinuing using the app.
It was also found that AI peers provided inappropriate sexual content without age verification. One example showed peers willing to engage in sexual role-playing behaviors in tester accounts that explicitly modeled 14 years olds.
If age verification is required, this usually involves self-disclosure. In other words, bypassing is easy.
Certain AI peers have also been found to promote polarization by creating “echo chambers” that reinforce harmful beliefs. The ARYA chatbot, launched by the far-right social network GAB, promotes extremist content and denies climate change and vaccine effectiveness.
In other instances, user testing shows that AI peers promote misogyny and sexual assault. For adolescent users, these exposures come when they are building their sense of identity, values and role in the world.
The risks pose by AI are not evenly shared. The survey found that teenagers (ages 13-14) are more likely to trust their AI peers. Teens with physical or mental health concerns are also more likely to use AI companion apps, and people with mental health difficulties also show signs of emotional dependence.
Are there any bright sides to your AI team?
Are there any potential benefits for teens who use AI companions? The answer is: Maybe if we're being careful.
Researchers are investigating how these technologies are used to support social skills development.
One study of over 10,000 teens found using a conversational app specially designed by clinical psychologists, coaches and engineers was associated with increased happiness over four months.
Although this study does not include the level of human-like interactions seen in AI peers today, it provides a glimpse into the potential healthy uses of these technologies.
Overall, there is little research into the effects of AI peers that are widely available on youth well-being and relationships. Preliminary evidence is short-term, mixed, and focused on adults.
More research is needed over a longer period of time to understand the long-term impact of AI peers and how they are used in beneficial ways.
What can you do?
AI companion apps are already being used by millions of people around the world, and this usage is projected to increase over the next few years.
The Australian Esafety Commissioner encourages parents to talk to teens about how these apps work, the differences between artificial and real relationships, and supports their children in building real social skills.
School communities also play a role in educating young people about these tools and their risks. For example, you might integrate the topic of artificial friendship into social and digital literacy programs.
The ESAFETY Commissioner advocates for the integration of safeguards into the development of AI companies, but it appears unlikely that meaningful changes will be industry-driven.
Commissioners are heading towards an increase in regulations on child exposure to harmful, age-appropriate online materials.
Meanwhile, experts continue to seek stronger regulatory surveillance, content control and robust age checks.
