Artificial intelligence (AI) has already been integrated into our society in a variety of areas. Voice control platforms penetrate more than 50 million homes across the US (Kelly, 2018). Children are now growing up with these devices as everyday tools. The devices have been developed to be more intuitive and engaging across all ages. Kelly describes her son recognizing the role and usefulness of their at-home Alexa device as early as age two. As children grow and develop with these devices, they attempt to put the devices into a conceptual framework they can understand, often placing them in a space as “…technology, but they also see these gadgets in psychological terms — as having emotions, as being capable of thought and friendship, and deserving of moral treatment (2018)” stated, University of Montana child psychologist, Rachel Severson. She goes on to say that as AI becomes increasingly sophisticated and “real,” individuals may form deeper attachments to AI devices, blurring the roles they play in our lives.
As AI chatbots increase in popularity and sophistication, from ChatGPT, Gemini, and Meta AI, their use of answering simple questions or assisting with research morphs into deeper forms of companionship. As of 2025, around 33% of teens are using chatbots for some sort of companionship (Robb). Forcing the possibility that minors have their first forms of mature relationships with chatbots rather than their peers. AI tools introduced at young ages, while children’s brains are developing and easily influenced, bring additional danger to the interactions children have with chatbots. There are also privacy concerns as parties involved in “conversations” with chatbots have their data saved to the companies’ servers. These can be intimate thoughts, feelings, and possible images while interacting with a “trusted” chatbot. Though such scenarios may sound extreme, reports of teens committing suicide after receiving it as a logical solution from chatbots have come to fruition (Chatterjee, 2025).
This begs the question of how we can prevent such harsh effects? As librarians, one of our core tenets is to allow access to knowledge. I do believe it is our responsibility to provide resources to the public about how this technology is impactful and how we can safely interact with it. Some options include: hosting open discussions, lectures, q-and-a’s, choosing an AI-centered book for a book club, and creating a well-rounded book display with additional resources. Hosting information sessions for parents on how to safely implement and monitor technology use, or how to talk to their teens about the dangers of such tech, are additional options.
On a grander scale, it is a social responsibility of the tech companies creating and distributing such technology to put in safety measures that prioritize the wellness of their users over the profit of their investors. Such as advice that offers professional help or seeking trusted adults if thoughts of self-harm occur, or alerting authorities if a minor is thought to be in danger.
Governments are considering legislation to combat the precarious scenarios chatbots may put children in. Michigan’s Senate recently proposed Bills 757-760 to address how AI chatbots interact with minors. Additional legislation has been floated to set a minimum age for social media, potentially following Australia’s lead earlier in 2026, which raised its minimum age to 16 for social media sites (Livingstone, 2026). However, since their implementation, they have had difficulties enforcing the ban, and it is being contested in the Australian Supreme Court.
The mental health effects of such technology are incredibly new, and we are learning every day how to adapt to its effects. It is a huge learning curve. As we see harmful effects across ages, geographical, and socioeconomic backgrounds, it is important to act in the interest of safety and care over monetary profit, which is a concept many companies appear to have trouble grasping. As educators and facilitators of knowledge, we should provide safe havens and various avenues of resources for the public that may not know how to tackle such new and daunting horizons.
Briney, J. (Host). (2026, May 31). CD339: Raising Screen Zombies (Episode 339) [Audio podcast episode]. In Congressional Dish. https://congressionaldish.com/cd339-raising-screen-zombies/
Chatterjee, R. (2025, September 19). Their teenage sons died by suicide. Now, they are sounding an alarm about AI chatbots. NPR. https://www.npr.org/sections/shots-health-news/2025/09/19/nx-s1-5545749/ai-chatbots-safety-openai-meta-characterai-teens-suicide
Kelly, S. (2018, October 16). Growing up with Alexa: A child’s relationship with Amazon’s voice assistant. CNN. https://www.cnn.com/2018/10/16/tech/alexa-child-development/index.html
Livingstone, H. (2026, January 22). Australia has banned social media for kids under 16. How does it work? BBC. https://www.bbc.com/news/articles/cwyp9d3ddqyo
Pappini, A. (2023, January 27). ChatGPT: A Library Perspective. Bryant University. https://library.bryant.edu/chatgpt-library-perspective
Robb, M.B., & Mann, S. (2025). Talk, trust, and trade-offs: How and why teens use AI companions. Common Sense Media. https://www.commonsensemedia.org/sites/default/files/research/report/talk-trust-and-trade-offs_2025_web.pdf
Hi Krystal, you bring up great points on the implications of AI companionship and mental health. This makes me wonder how such chatbots affect users’ interactions with other people in the long run, especially in younger users—what will they do when they meet people who disagree with them or don’t do what they say? I don’t believe we should ban minors from using certain technologies/platforms though, because that would be a violation of not only our intellectual freedoms but also our privacy, but I do wonder if regulating how AI is designed and marketed would help, i.e., making them less human-like. Though we might understand that chatbots are LLMs and therefore don’t “think” and “feel” like real people, this might not be obvious to everyone, especially those who are less tech-literate. Friendly, human-like chatbots technically may fulfill the “delight” heuristic of design when they serve as companions, but during this “loneliness epidemic,” we should be building human connections, not obsequious technology.
Hi Anita, yes, banning minors of such platforms proves a sticking point especially when considering our First Amendment right here in the US. Adding further regulations to AI seems like a logical solution, but has already faced push back in the current US congress, due to it “stifling” innovation. Since tech-companies creating and releasing such chatbots, have not created proper safeguards, it does to appear to be at an impasse, since the companies cannot be trusted to put safety first, and the US government has so far lacks the will to regulate.
In addition, to properly educating ourselves and the public about the benefits and potential dangers of AI chatbots, I do think the focus on building human connections is a great way to combat the rise of companion chatbots. Creating a human connection in reality will prompt a person to seek out a voice they know, rather than a tech they don’t, or at least provide them with the option.
Hi Krystal — it’s truly really scary what AI chatbots sometimes say in response to emotional conversations. I know of an acquaintance who was using Chat GPT as a therapist — and while they were experiencing a difficult time in their relationship with their partner, they turned to Chat GPT for advice on whether or not to part ways. I think, internally, they already knew what their decision was, so subconsciously, they were “feeding” Chat GPT that information. Eventually, Chat GPT did make the conclusion to break up with their partner. Talk about an echo chamber! As information professionals, I think it is indeed our job to, in your words, teach the public “how this technology is impactful and how we can safely interact with it…”
Hi Tina, I also have a good amount of friends who use Chat GPT as a therapist. When we discuss the platform they are aware it is essentially an echo chamber, as you put it, and not necessarily anything that will push them to make a change they don’t want to do, even if it is for the better, as an actual therapist might. They know it is just regurgitating what it is fed by them, yet, they still look refer to it as if for a sense of validation in their decision making. For reference, most of those friends range from the ages of 24-30. Well-formed adults, who are capable to make well-thought out decisions without a chatbot. YET, they still check-in with it, even if they don’t let it rule their decision making.
How younger people, especially teens may be rely on the tech is scary, especially since the tech has proven to not have the appropriate safeguards to ward away self-harm, or lude behavior. It has proven the opposite, egging on such behavior, if that appears to be the “logical” solution based on information it is fed. Educating ourselves and the public on the pitfalls, not just the benefits, of this kind of technology is essential to making sure it used appropriately.
Krystal – I appreciate your thoughts on this complex topic and I’m so glad you brought up the idea of young people using LLMs for companionship. This statement resonates: As we see harmful effects across ages, geographical, and socioeconomic backgrounds, it is important to act in the interest of safety and care over monetary profit, which is a concept many companies appear to have trouble grasping.