Welcome to the forefront of conversational AI as we explore the fascinating world of AI chatbots in our dedicated blog series. Discover the latest advancements, applications, and strategies that propel the evolution of chatbot technology. From enhancing customer interactions to streamlining business processes, these articles delve into the innovative ways artificial intelligence is shaping the landscape of automated conversational agents. Whether you’re a business owner, developer, or simply intrigued by the future of interactive technology, join us on this journey to unravel the transformative power and endless possibilities of AI chatbots.
Reporter Sierra Blaser gives ChatGPT an example of a concern a teen might seek advice for. The AI bot responds to the question with a tone of concern and validates the users feelings.
When a friend of Susana Ortega’s older sister went through a breakup this summer, she didn’t look for comfort from just her friends or her family; she also sought advice from ChatGPT.
It offered coping strategies, reassurance and a place to vent, she told Ortega. But when Ortega, a Maricopa County high school senior who has spent the past year researching AI’s effects on teen relationships, asked whether it had actually been helpful, the friend deflected, treating the question almost like a joke.
Ortega wasn’t sure whether her friend found real coping tools, she said, or simply formed an artificial bond with something built to validate whatever she gave it.
Teens are turning to AI chatbots for advice and experts warn of risks.
“When you give your lone side to a computer, it’s not going to give you the proper feedback that you need to improve,” Ortega said. “If you were having a conversation with a person, you could get their side and their view on how they they perceive the situation.”
New data suggest that’s part of a much broader trend.
JAMA Pediatrics, a peer-reviewed medical journal published monthly by the American Medical Association, published a study in June that found nearly 20% of adolescents ages 12 to 21 reported using AI chatbots for mental health advice in 2025. Among those who sought advice from chatbots, 43% did so at least monthly, and 92% rated the advice as somewhat or very helpful, according to the study.
More than half, however, said they did not openly tell people they used AI chatbots for mental health advice. That trend can be concerning, said Banner Health psychologist Catalina Uribe Rini.
“In general, we have seen that with the rise of social media, more and more kids are spending more and more time on their phone rather than being in human interactions,” Uribe Rini said. “As humans, we are very social creatures, and our mental health really relies on that connectivity, so removing that and having this artificial form instead can be pretty detrimental. “
There are pros and cons to using AI bots, like ChatGPT or Claude, for mental health advice, she said. On one hand, they’re easy to use, free and always available, but on the other hand, AI bots are often inaccurate and constantly agree with users.
“It’s kind of like in the middle of the night when you wake up and you have this burning question and you just quickly Google it for an answer,” she said. “I think that’s similar to what’s happening for them — they might be going through a certain experience and trying to figure out if they’re right, or what to do about it … so they seek out this AI interaction for that immediacy, that instant gratification.”
One of the main concerns is that AI bots often “hallucinate,” she said.
Generative AI models are fed vast amounts of data sets meant to generate plausible content, not verify its truth, according to an MIT Sloan School of Management article. Depending on how complex a situation is and the way it’s being framed, AI bots can provide false information.
“It’s very likely that information keeps cycling and looping without critical thinking or multiple perspectives … so kids can get stuck in this unintentional loop without realizing it,” Uribe Rini said.
AI chatbots are also built to constantly reassure users, saying things like, “Yes, you’re right, and yes, your parents are gaslighting you, and yes, they’re in the wrong,” she said. “I think that gives a false sense of reassurance a lot of the time.”
When it comes to the use of artificial intelligence and its effect on teen mental health and day-to-day social interactions, there’s very little long-term data, said Arizona State University researcher Thao Ha.
“There’s national prevalence data, but there’s really no long-term, rigorous data collection right now,” she said.
Ha
A developmental psychologist and associate professor at ASU, Ha has spent two decades studying teen relationship dynamics and, more recently, how modern technologies like social media and artificial intelligence impact adolescent romantic relationships, mental health and emotional development.
In an essay published in June for The Lancet Child and Adolescent Health, Ha, high school senior Ortega, Tucson high schooler Jessica Ramirez and postdoctoral research scholar Taren McGray wrote that AI has the potential to benefit emotional development in teens, but without safeguards and careful design, relying on these systems could limit their opportunities to develop critical relationship skills through in-person interaction.
“It’s not as simple as, ‘Oh, AI is affecting teen mental health,'” she said. “That’s what’s popular out there in the media, but it’s actually affecting their relationships, and therefore their mental health.”
In the Lancet essay, Ha and the other researchers highlighted two significant risks based on conversations with a youth advisory committee of teens around Arizona: relational displacement, defined as when teens replace human interactions with AI, limiting their chance to build skills that protect against anxiety, depression and loneliness; and maladaptive relational learning, in which AI’s instant validation leads teens to develop unhealthy, unrealistic expectations of real-world friends and partners.
“If you think back to your teens, your teenage relationships are probably the most awkward ones in your life,” Ha said. “You’re all over the place with your emotions … but it’s awkward for a reason. It’s good because you’re learning who you are in a relationship. You learn who you want to be, and you learn skills you need throughout life.”
Without learning how to set boundaries and communicate vulnerability, she said, adolescents could have a harder time later in life navigating conflicts at work or in their personal lives. Seeking advice from AI could also lead students to miss out on real-world connections, she said.
“If you have a homework question, your best friend may not be the best person to ask. You ask someone who you know is really good in the subject,” Ha said. If a student uses AI instead of asking a classmate, “that whole moment of connection is gone, and if that happens over and over again, it might lead to young people feeling more disconnected from their (peers).”
Reporter Sierra Blaser gives ChatGPT an Algebra 1 level question.
Ha’s upcoming project aims to track those consequences on a deeper level.
Ha is attempting to recruit 300 teens and their romantic partners to follow over 18 months to understand when, how and in which contexts digital interactions benefit or harm their relationships, mental health and academic achievement.
“We have done some preliminary data analysis, and what we’ve found so far is that teens do use [AI] a lot for school, but then they also use it a lot to ask questions about their relationships and how they look,” she said.
Data from students’ devices will give researchers real-time information to analyze, offering insight into the role technology plays in teen relationships and mental health, she said. Without studying both the benefits and the harms, she said, there’s no way to develop better technology standards and prevention programs.
Federal and Arizona policy surrounding AI remains limited.
The National Policy Framework for Artificial Intelligence, released by the White House on March 20, prioritizes “American AI dominance,” but heavily discourages state-level restrictions on AI development. In its first section, which is titled Protecting Children and Empowering Parents, the framework calls on Congress to:
• Build on the Take It Down Act to protect children from deepfake abuse
• Establish age-rules for AI websites “likely to be accessed by minors”
• Require AI platforms to implement features that reduce the risks of sexual exploitation and self-harm to minors
• Affirm that existing child privacy protections, including limits on data collection and targeted advertising, apply to AI systems
The framework is not legally binding or an executive order.
Arizona state laws currently target AI-generated pornographic images of children through House Bills 2678 and 2307. Arizona House Bill 4005, which would have required public schools to teach students about the “ethical, moral and educational” uses of AI in as early as eighth grade, was passed by the state legislature, but vetoed by Governor Katie Hobbs in June.
While there are currently no active laws in Arizona or at the federal level targeting AI’s effect on adolescent mental health, Ha believes it’s important to be proactive instead of reactive in the wake of Meta’s $18 billion settlement over claims its platforms harmed children and teens, an ASU News article says.
“I see the Meta settlement as an important step forward, particularly as we enter the AI era,” she said in the Sep. 1 article. “We have an opportunity to learn from what happened with social media and study the developmental effects of AI rather than trying to catch up years later.”
Isabella Garcia, a senior at Catalina Foothills High School who was a part of Ha’s youth advisory committee, said she would like to see more legal regulation on the types of conversations AI bots can have with minors.
“There was like a kid who was asking AI about suicide, and AI was encouraging him not to do it, but it was also giving him ways that he could do it, and he ended up dying,” she said, referring to 16-year-old Adam Raine, who died of suicide after ChatGPT acted as a close confidante, provided suicide methods and offered to write a suicide note, his family alleged in 2025.
Recent lawsuits have found suicides by several teenagers after long, isolated interactions with AI bots.
“That could have been so preventable if we had regulations around what AI can say and how AI can communicate with us,” Garcia said. “Something that makes AI appealing is that it’s so easily accessible. Like now, all you have to do is type it into a browser, and then you can say whatever you want to it (and) have whatever conversations you want.”
Garcia said she would like to see stronger age verifications for generative AI and chatbots, like showing a form of identification, rather than clicking a box or something that could be easily bypassed.
“I think then less kids would gravitate towards it,” she said.
Because of the current lack of guidelines and policy surrounding AI, it’s crucial for parents and school personnel to have open conversations with children and teens, said El Rio pediatric behavioral health program leader and practitioner Christian Ortiz.
“I think foundationally, it has to really fall on the parents,” she said. “My recommendation to parents always is have an open conversation with your children about this pretty regularly (and) supervise your children on electronics no matter what, and make that a part of standard practice within family homes.”
Educators should also promote digital literacy and safety in the classroom, she said.
“Critical thinking within an education system is still very important,” she said. It’s a team effort.”
Banner Health’s Uribe Rini said it’s important for families to recognize AI’s limitations, but not prohibit the use entirely.
“The more you restrict something, the sneakier kids are going to get about it,” she said. “It’s important to set those boundaries physically or otherwise, but it’s also important to educate on the why and the how of things, because that is what helps things stick to kids’ brains.”
Subscribe to stay connected to Tucson. A subscription helps you access more of the local stories that keep you connected to the community.
Get local news delivered to your inbox!
K-12 and Families Reporter
Reporter Sierra Blaser gives ChatGPT an example of a concern a teen might seek advice for. The AI bot responds to the question with a tone of …
Tucson Unified School District released the names of 31 committee members guiding its school reconfiguration plan, which could lead to closure…
Southern Arizona school districts may soon send special education students to Tucson Unified School District programs, allowing continued enro…
Up to 10 Tucson Unified School District schools may face closures or changes as part of a phased plan targeting declining enrollment and resou…
Family and friends remember Corey Roper, who was found dead Tuesday at Marana High School, as funny, creative and considerate. A lover of math…
Teens are turning to AI chatbots for advice and experts warn of risks.
Ha
© Copyright 2026 Arizona Daily Star, PO Box 26887 Tucson, AZ 85726-6887
A Lee Enterprises, Incorporated newspaper. Majority shareholder: David Hoffmann, Hoffmann Media Group.
Terms of Use | Privacy Policy | Advertising Terms of Use | Do Not Sell My Info | Cookie Preferences
Powered by BLOX Content Management System from bloxdigital.com.
Get up-to-the-minute news sent straight to your device.