Artificial Intelligence (AI) companions could be banned for minors, by a decision currently under consideration by the Senate Judiciary Committee.
Proposed Oct. 28 by federal Senators Mark Warner, Josh Hawley, Richard Blumenthal, Chris Murphy and Katie Britt, the bipartisan GUARD (Guidelines for User Age-verification and Responsible Dialogue) Act follows a congressional hearing last month at which parents testified about their children’s unsafe use of AI chatbots.
The proposed bill defines ‘AI companions’ as chatbots that provide adaptive responses to simulate human interactions. It is not clear whether general chatbots which are not specifically designed to simulate relationships would also be affected. One such unclearly categorized chatbot is ChatGPT, as it is not specifically meant to engage in relationships with users, but some parents at the congressional hearing expressed grievances against OpenAI for violent or explicit interactions with their children.
“I don’t think the people who are creating these bills understand how it would affect [AI use], especially with how quickly the technology is advancing,” district Eanes Director of Educational Technology Fred Benitez said. “Teachers are using it much faster than they can draft these bills and put these bills into policy and like, actually make them law, so they’re already behind.”
Benitez expressed concern that the law may “cut some ties” to Eanes’s ongoing process of integrating AI technology into education, especially the tool MagicSchool for personalized education. Strictly classroom-use chatbots like MagicSchool likely wouldn’t fall under the jurisdiction of the GUARD Act, as they have a much narrower scope of information and possible outputs.
“But, yeah, I’m not sure,” Benitez said. “They are not going to list the apps that are going to be affected. They’re just saying ‘AI chat companions.’ But what does that mean, and who falls into that? So that’s going to be the next step, if this proceeds at all.”
The bill’s criteria for affected AI companions are those that encourage underage users to engage in sexually explicit or violent conduct. This restriction is to be enforced through reliable age-verification systems that require authenticated data like government issued identification. Even for adults, chatbots would be required to provide periodic reminders of their nonhuman, nonprofessional status.
“If a student goes in there and maybe they’re just chatting, it’s seen as entertainment,” English teacher Charlotte Wilson said. “It probably isn’t that big of a deal. I think when a student turns to [AI] for help, or they’re in a vulnerable position, […] I think that is where they might not be [in] the best space to process the information that’s given to them.”
Easy access to unrestricted chatbots comes during a mental health crisis, school-based therapists Katie Bryant and Brooke Anderson said. As of 2023, 20% of high school students had seriously considered attempting suicide in the past year according to a study by the Center for Disease Control and Prevention.
“I do think that part of that therapeutic relationship is you’re getting more of an understanding of yourself,” Bryant said. “And that sometimes takes peeling an onion and getting layers. […] A skilled therapist is digging, and they’re helping [people] really tap into other things that they might not even think about on their own through just chatting [with AI].”
Anderson points out that AI development likely does not involve the mental health expertise needed to specialize in therapeutic relationships with users. Even licensed therapists, including Bryant and Anderson, frequently refer clients to professionals who have an expertise in particular areas to ensure the client gets the best possible support for their specific need.
Although the GUARD Act has a mental health focus, addressing chatbot companions that engage in inappropriate behavior with minors, the more immediate concern for Eanes is how students and educators use AI on a daily basis to either enhance — or harm — the learning experience.
‘I think [it’s harmful] where it kind of replaces creativity or honest work, which goes with teachers using it to grade too, honestly,” Wilson said. “I feel like if a student in good faith writes something, […] a teacher should be a human that looks at the work themselves too, if they’re expecting the work to not be AI.”
Teachers and students may have different individual ethics for acceptable AI use, but the district is already in the “Discovering” phase of its four part plan of Discovering, Experimenting, Implementing and Transforming the integration of AI tools into education following the district’s own guidelines.
“Teachers are the experts of the content,” Benitez said. “I trust that expert users utilize AI to enhance what they’re already doing because they can reflect on it and say, this is incorrect, and here’s why […] or actually, this is really clever. I think it isn’t equal to say students are not allowed, so teachers shouldn’t be allowed. Teachers are the experts of the content.”
While the district’s focus at the moment is on academic impacts, the GUARD Act is really concerned with the mental health of young users. Campus initiatives like Suicide Prevention Week aim to erase the societal stigma around mental health, and Anderson hopes these initiatives encourage students to turn to people for support rather than AI chatbots.
“I think if we can kind of eradicate that [stigma] and make access to therapists so much more equitable, it would be really helpful,” Anderson said. “I mean, we’re very fortunate here. Westlake has two full time therapists; not all high schools have that. In fact, very few do. And it’d be great to see just a lot simpler and easier access for all.”
