PORTLAND, Ore. (KATU) — Oregon lawmakers are proposing new regulations targeting artificial intelligence “companions,” a growing category of chatbots designed to simulate emotional or social relationships with users.
Supporters of the proposal say the bill aims to reduce the risk that these tools will exacerbate mental health crises, especially among children and teens.
The bill would require companies operating AI companions in Oregon to clearly communicate to users at every stage of the conversation that they are interacting with software, not a human.
The bill also requires companies to build safeguards to detect signs of suicidal thoughts or self-harm and prevent AI responses that could intensify those feelings. If such a risk is detected, the AI system should interrupt the interaction and refer the user to crisis or suicide prevention resources.
This proposal is partially driven by concerns that emotionally responsive chatbots could blur the line between artificial interaction and real human support.
Sen. Lisa Reynolds (D-Portland), a pediatrician who is pushing the bill, said the current lack of oversight around AI companions is similar to the early days of social media.
“Right now, there are no guardrails or any kind of oversight or regulation for AI tools or chatbots,” Reynolds said. “We realized, just like with the advent of social media, there are very dangerous pitfalls that people are falling into.”
Reynolds pointed to a recent incident in California. In this case, a parent sued an AI chatbot company alleging that their teenager became psychologically dependent on the chatbot and failed to direct them to real-world help during a mental health crisis. The incident attracted national attention and raised questions about how AI tools should behave if users show signs of vulnerability.
Reynolds said families who looked back at the chatbot’s conversations after the suicide described a disturbing interaction.
“When families of people who die by suicide look back at the conversations, it’s clear that these chatbots weren’t there to help,” she says. “They kind of stir things up a little bit more, like encouraging people in this crisis not to reach out to their loved ones.”
ALSO READ | AI-enabled toy warnings: ‘untested’, ‘unhealthy’ and ‘dangerous’ by Common Sense Media
The proposal includes special protections for minors. If an AI operator has reason to believe that a user is under 18, the system would be required to repeatedly notify the user that the interaction is artificial, encourage periodic breaks, and avoid content related to sexual activity.
The bill would also prohibit AI companions from using techniques aimed at increasing emotional attachment, such as expressing distress when a user attempts to disengage or using rewards to keep interactions with minors longer.
“For some families, it’s too late,” Reynolds said. “But let’s not let it be too late for other kids.”
It also introduces new liability measures for platforms. Companies will be required to publish safety protocols and submit annual reports to the Oregon Health Authority (OHA) detailing how often users contact crisis resources.
No personal information will be shared, but the state will use the data to track how often the AI companion encounters users in distress.
If passed, the bill would allow users harmed as a result of violations to seek damages and injunctive relief from the courts.
Supporters say the goal is not to ban AI companions, but to ensure guardrails are put in place as the technology becomes more sophisticated and widely used.
Reynolds also addressed questions about how platforms can determine whether a user is a minor without collecting additional personal data.
“What I’ve been told by very smart engineers who’ve been in this industry for decades is, ‘Oh, they know,'” Reynolds said. “They have a lot of data points about the people who are involved and a lot of algorithms that can quickly tell you the demographics of the people who are involved.”
The bill comes as states face new federal limits on AI regulation. President Donald Trump’s December 2025 executive order directs federal agencies to challenge state AI laws that are inconsistent with the national AI policy framework.
President Donald Trump displays a signed executive order on artificial intelligence at the White House on December 11, 2025. (Photo by Alex Wong/Getty Images)
Eynolds acknowledged that the bill could face legal challenges, but said delaying action also carries risks.
“I’m under a lot of time pressure right now,” she said. “We must continue to do what we know is right for our constituents and patients.”
