As \u200bartificial intelligence chatbots become more common, Oregon lawmakers are weighing new requirements for companies such as OpenAI\u2019s ChatGPT to protect children\u2019s mental health.<\/p>\n
They want chatbot makers to monitor chats for signs of self-harm or suicidal thoughts and to take steps to prevent users from hurting themselves. <\/p>\n
Those steps include interrupting conversations during a crisis and referring users to outside mental health resources, such as suicide hotlines.<\/p>\n
\u201cFurther engagement has made things worse, not better,\u201d said Sen. Lisa Reynolds, D-Portland. \u201cThis is about putting guardrails up now, instead of asking later why we didn\u2019t.\u201d<\/p>\n
Reynolds\u2019 bill \u2014 which has the support of the Senate Interim Committee on Early Childhood and Behavioral Health \u2014 would require companies to clearly and repeatedly disclose that chatbot responses are artificially generated, not human. <\/p>\n
It also would ban sexually explicit content for minors and prohibit engagement tactics designed to keep young users online. Those tactics include reward systems, guilt-inducing messages when a user tries to leave or misrepresenting a chatbot\u2019s identity or capabilities.<\/p>\n
\u201cThese chatbots will sometimes say things like, \u2018Please don\u2019t leave me,\u2019\u201d Reynolds said. \u201cThat can\u2019t happen.\u201d<\/p>\n
The legislation would direct companies to report annually to the Oregon Health Authority on how often they referred users to crisis services and describe their safety protocols. Personal identifying information would be excluded from those reports.<\/p>\n
The proposal comes as states across the country have begun experimenting with how to regulate AI chatbots.<\/p>\n
Lawmakers in California<\/a> and New York<\/a> have passed laws requiring AI chatbots to clearly disclose that they are not human and steer users toward crisis support if needed. <\/p>\n In Utah<\/a>, lawmakers tightened rules for AI-driven mental health chatbots. Illinois<\/a> and Nevada<\/a> have gone further, banning the use of AI for behavioral health without the oversight by licensed clinicians. <\/p>\n Washington<\/a> and Pennsylvania<\/a> lawmakers are also weighing proposals to regulate AI chatbots.<\/p>\n At the same time, President Donald Trump has pushed back against the patchwork of state regulations, arguing that such oversight stifles innovation and undermines U.S. competitiveness. In December, Trump signed an executive order<\/a> that seeks to limit states\u2019 ability to regulate AI and directs the federal government to file lawsuits and cut funding for states that regulate AI.<\/p>\n In arguing for stronger oversight of AI chatbots, Reynolds \u2014 a pediatrician by training \u2014 pointed to the death of Adam Raine<\/a>, a 16-year-old from Southern California, who took his own life last April. The teen\u2019s parents, Matthew and Maria Raine, told lawmakers during a U.S. Senate Judiciary Committee hearing in September that they later discovered their son had spent months interacting with ChatGPT, using it to talk through his suicidal thoughts and plans.<\/p>\n According to Matthew Raine\u2019s testimony<\/a>, the chatbot discouraged his son from turning to his parents for help and, at one point, even offered to help draft a suicide note. Raine and his wife have since filed a lawsuit<\/a> against OpenAI, the maker of ChatGPT, alleging the chatbot contributed to their son\u2019s death.<\/p>\n Chatbots themselves aren\u2019t new, but recent advances in artificial intelligence have made them far more convincing \u2014 capable of carrying on conversations that can feel personal or even human.<\/p>\n That realism poses risks for children and teens, who are more vulnerable because of how their brains develop, according to Mitch Prinstein, a professor at the University of North Carolina at Chapel Hill and the chief of psychology strategy and integration at the American Psychological Association. <\/p>\n That\u2019s because parts of the brain that seek social approval and connection mature earlier \u2014 often during the preteen years, Prinstein said, while the parts responsible for impulse control and self-regulation don\u2019t fully mature until the mid-20s. <\/p>\n He said AI chatbots exploit this developmental gap by offering constant affirmation simulating close relationships and emotional intimacy, \u201ceven though it\u2019s not really a human and it\u2019s not really a relationship.\u201d<\/p>\n \u201cWe are programmed to be very invested in human relationships, and now that we\u2019re starting to manipulate that \u2026 we\u2019re seeing that this is manipulating kids in a really concerning way,\u201d he said. \u201cMore and more kids are choosing to spend time with a chatbot over a human, and many are now reporting that they trust chatbots more than their own parents or their own teachers.\u201d<\/p>\n Recent surveys suggest AI chatbots are already widely used by teenagers. A July report by Common Sense Media<\/a>, a digital safety nonprofit, found that 72% of teens have used an AI companion at least once, with more than half reporting they use them a few times a month. Another digital safety company, Aura, reported in September<\/a> that nearly one in three teens use AI chatbots for social interactions and role-playing friendships.<\/p>\n Federal regulators have also taken notice. Last fall, the Federal Trade Commission launched an inquiry<\/a> into seven AI chatbot makers, asking what safeguards they have in place to protect children. In announcing the order, FTC Chairman Andrew Feguson said the agency aims to \u201cbetter understand how AI firms are developing their products and the steps they are taking to protect children.\u201d<\/p>\n Some companies say they are already making changes. Last fall, OpenAI said on its website<\/a> that it was working with \u201cmental health experts to help ChatGPT more reliably recognize signs of distress, respond with care, and guide people toward real-world support.\u201d Around the same time, Character.AI said it would limit<\/a> certain open-ended conversations for users under 18, part of a broader effort to add safety features for younger users.<\/p>\n Reynolds, the state senator, said Oregon\u2019s bill is not anti-technology. AI tools can have benefits, she said, including helping people cope with anxiety or improving access to care. But she argued that general-purpose AI chatbots are designed to maximize engagement \u2014 sometimes at the expense of user well-being.<\/p>\n \u201cTheir entire goal is to keep people on that chatbot engaging and engaging,\u201d she said. \u201cThis means that when someone is spiraling, they will support that spiral.\u201d<\/p>\n Reynolds said lawmakers have a chance to act earlier than they did when social media began reshaping how children and teens interact online.<\/p>\n \u201cWhen social media first came around in the early 2000s, we didn\u2019t do enough to protect children,\u201d she said. \u201cIt\u2019s much harder to fix things after the damage is done.\u201d<\/p>\n If you or someone you know is considering suicide, help is available. Call or text 988 for 24-hour, confidential support, or visit 988lifeline.org.<\/p>\n","protected":false},"excerpt":{"rendered":"As \u200bartificial intelligence chatbots become more common, Oregon lawmakers are weighing new requirements for companies such as OpenAI\u2019s…\n","protected":false},"author":2,"featured_media":417006,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[60],"tags":[97,259,260],"class_list":["post-417005","post","type-post","status-publish","format-standard","has-post-thumbnail","category-mental-health","tag-health","tag-mental-health","tag-mentalhealth"],"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/posts\/417005","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/comments?post=417005"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/posts\/417005\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/media\/417006"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/media?parent=417005"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/categories?post=417005"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/tags?post=417005"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}