Senate examines AI companions for minors, seeks industry input on safeguards
An interim study explored age verification, parental oversight, and platform accountability as lawmakers prepare to revisit artificial intelligence legislation.
Oklahoma senators examined potential safeguards for children using artificial intelligence companions Sept. 14, with the committee chair urging businesses to help shape legislative language before the next session.
The Senate Technology and Telecommunications Committee held the study at the request of Sen. Warren Hamilton, R-McCurtain. The agenda for Interim Study 2026-44 listed cybersecurity specialist Ryan Montgomery, founder and CEO of Pentester.com, and Jay Paul Gumm, communications director for the Oklahoma Institute for Child Advocacy.
Committee Chair Ally Seifried, R-Claremore, said lawmakers face the challenge of protecting minors while preserving beneficial uses of AI. She pointed to everyday applications, including music recommendations and food-delivery suggestions, as examples of services she wants to continue.
“Your time to work with us is now,” Seifried told industry representatives and businesses concerned about potential legislation. She said technical wording complicated work on an AI bill during the spring session and encouraged companies to engage with her and Hamilton during the interim.
The discussion during the study extended beyond companion chatbots to AI-generated imagery, online exploitation, and the responsibilities of companies developing or hosting AI tools.
Montgomery described concerns about minors receiving inappropriate or harmful responses from conversational systems. He discussed requiring identification to verify users’ ages and suggested that verified parents could establish accounts for their children.
Sen. Carri Hicks questioned whether a parent-linked account system could be misused by adults who falsely claim responsibility for a child. The exchange highlighted unresolved questions about how verification would work and whether it would prevent exploitation.
Montgomery also called for stronger safeguards tailored to children and additional cybersecurity resources for agencies investigating crimes against minors.
Gumm focused on accountability for developers and platforms involved in generating or distributing sexually explicit depictions of minors. He advocated legal remedies allowing victims to seek damages and court-ordered removal of harmful content.
He also said legislation should distinguish legitimate, general-purpose AI development operating within safety protocols from entities that knowingly or recklessly facilitate harmful uses. His presentation emphasized parental involvement alongside potential enforcement measures.
The study follows work on Hamilton’s Senate Bill 1521 during the 2026 session. The Senate passed its version 43-0 in March, and the House passed an amended version 90-0 in April. The bill returned to the Senate but did not complete the legislative process.
According to the Senate’s summary of the House-amended version, the measure would have required disclosures telling minors they were interacting with AI, restricted rewards intended to encourage continued interaction, and required protocols directing users expressing suicidal thoughts to crisis services. It also proposed attorney general enforcement and civil penalties. Those provisions were proposals, not requirements enacted through the bill.
The committee adjourned the interim study without voting on legislation. Seifried said work on the issue would continue ahead of the 2027 session.
