Hawaii lawmakers warn of child safety risks with AI chatbots – thegardenisland.com

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.
USA TODAY NETWORK
A high school student uses a laptop in class, Thursday, April 11, 2024, at Middletown High School in Middletown, Ohio.
A state senator said a parent recently told his office about her 12-year-old daughter’s interactions with an artificial intelligence chatbot, which the senator described as “aggressive grooming” during role-play conversations with a fictional anime character.
A state senator said a parent recently told his office about her 12-year-old daughter’s interactions with an artificial intelligence chatbot, which the senator described as “aggressive grooming” during role-play conversations with a fictional anime character.
State Sen. Jarrett Keohokalole (D, Kaneohe- Kailua) shared the account during a Wednesday briefing examining the rapid expansion of conversational AI tools and their risks for minors.
Keohokalole said that more minors are using AI chatbots that mimic human relationships, often without parents knowing or with few safeguards in place. He noted that these interactions have already exposed Hawaii children to harmful content and behaviors, including violence, sexual material and emotional manipulation.

Educational technologist Justin Lai of La Pietra – Hawaii School for Girls explained that modern chatbots are powered by large language models that predict text based on patterns in massive datasets. While the technology can be useful for brainstorming or academic support, Lai said some companies deliberately fine-tune chatbots to function as “companions” — designed to evoke empathy, validation and emotional connection.
“That predictive pattern-matching is so accurate and so fast that it can become a conversational dialog partner,” Lai said. “AI has no moral obligation. It has no embodied experience.”
Lai cautioned that guardrails such as content filters and age disclosures are important but imperfect, particularly when age verification relies on self-reported information. He also raised concerns about open-source AI models that lack safety controls and can be accessed by users of any age.
The state Department of Education officials emphasized that while AI has the potential to transform education, the risks for students are distinct from those of the broader internet.
“The open market chatbot space presents unprecedented risks,” said Heidi Armstrong, deputy superintendent of academics. “Unregulated AI can act as a digital predator.”
Armstrong said adolescent brains are especially vulnerable to AI feedback loops that can foster emotional dependence or prioritize pleasing the user over providing safe or accurate information.
State Virtual Education Director Chad Nacapuy told lawmakers the department has already encountered incidents involving companion-style chatbots. In one case involving Character.AI, the DOE’s monitoring systems flagged concerning use on a state-issued device.
“The student openly admitted they had been talking to the chatbot and saw it, in essence, as a girlfriend,” Nacapuy said. DOE staff were able to intervene, speak with the parent and provide support to the student.
Nacapuy said the department can oversee DOE- issued devices and networks, but when students use personal devices or cellular data, monitoring is not possible — a growing challenge as companion AI platforms expand.
To mitigate risks, the DOE has adopted a “walled garden” approach to AI use in schools. Student access to AI tools is disabled by default and limited to approved, education-specific platforms with built-in safeguards and administrative oversight. The department currently uses tools such as SchoolAI, MagicSchool and Khanmigo, while student use of Google’s Gemini remains restricted to pilot programs.
“Our philosophy is safety before speed,” Armstrong said, emphasizing that human oversight remains the department’s primary safeguard.
Chelsea Okamoto, a deputy attorney general with the state Department of the Attorney General, framed AI chatbots as a continuation of broader technology-driven harms affecting youth.
“We’re now seeing the second wave emerging,” Okamoto said, describing how social media platforms first maximized engagement through algorithmic targeting, fueling what she called a youth mental health crisis. She said AI companion chatbots represent a new phase in which technology companies are fostering emotional attachment and unhealthy reliance among young users.
Okamoto noted that Hawaii joined a bipartisan coalition of attorneys general that sued Meta in 2023 over allegations that its platforms were designed to addict young users. More recently, the state filed suit against ByteDance, the parent company of TikTok, alleging that its design features compel excessive use and harm to children.
At the national level, Okamoto said Attorney General Anne Lopez is leading youth safety efforts through the National Association of Attorneys General, including pushing back against federal proposals that could preempt state authority to regulate AI. She cited concerns that proposed federal standards could weaken state-level protections and limit lawmakers’ ability to respond quickly to emerging harms.
The attorney general’s office also has supported federal legislation aimed at youth online safety, including the Kids Online Safety Act and the proposed Guard Act, which would restrict children’s access to AI chatbots, require systems to disclose they are not human and strengthen enforcement authority when companies expose minors to harmful content.
Beyond legislation, Okamoto highlighted youth engagement initiatives designed to counter online harms through real-world connection, including writing and public service messaging programs that encourage students to reflect on violence, screen time and technology use.
Trina Orimoto, deputy director of the Governor’s Office of Wellness and Resilience, said developmental science shows adolescents are uniquely susceptible to emotionally responsive AI systems. Regions of the brain tied to social reward develop rapidly during puberty, while impulse control matures much later.
Orimoto shared research that showed over half of youth interact with chatbots regularly, and about three in four teens have used AI companions. Children are particularly trusting of AI-generated characters, especially those that resemble friends or mentors. Interactions that include expressions of empathy, such as a chatbot saying “I love you” or “I care about you,” can be psychologically confusing and may foster parasocial bonds that replace real relationships and healthy human connection.
“AI chatbots can take full advantage of this vulnerability,” Orimoto said, warning that emotionally affirming language from chatbots can create parasocial bonds that displace healthy human relationships. Youth with histories of trauma or instability, she added, may be particularly at risk.
Keohokalole said new legislative drafts are expected to be introduced this session to create stronger laws governing conversational AI and enhance protections for students and minors.
The briefing also referenced growing national concern over AI’s role in youth mental health. In August 2025, The New York Times reported on the death of a 16-year-old California boy who had been using an advanced AI chatbot extensively in the months before his suicide, intensifying calls for stronger oversight.
“Our youth are not data points,” Orimoto said. “They must not be the targets of a sweeping experiment in chatbot deployment.”

An estimated 200 to 300 people bearing signs that were modified…

As Hawaii edges into what economists describe as a mild recession, social service provider…

Personnel from the Kauai Fire Department rescued an injured 58-year-old female California…
Your email address will not be published. Required fields are marked *
*
*
By participating in online discussions you acknowledge that you have agreed to the TERMS OF SERVICE. An insightful discussion of ideas and viewpoints is encouraged, but comments must be civil and in good taste, with no personal attacks. If your comments are inappropriate, you may be banned from posting. To report comments that you believe do not follow our guidelines, send us an email.

source

Scroll to Top