A Senate subcommittee heard parents and experts on chatbot harm

What happened

On 16 September 2025, the Senate Judiciary Committee's Subcommittee on Crime and Counterterrorism held a hearing titled Examining the Harm of AI Chatbots. The committee's own hearing page lists five witnesses: two parents identified as Megan Garcia and Matthew Raine, a mother identified only as Jane Doe, Robbie Torney of Common Sense Media, and Mitchell Prinstein of the American Psychological Association. Their written testimony is posted on the same page. A hearing record documents what was said and asked; it is not itself a law or a finding of fact.

What the documents show

Prinstein's written testimony states he appeared on behalf of the American Psychological Association, representing what he describes as over 173,000 psychologists, students and affiliates, and frames AI chatbots as tools whose effects on cognition, emotion and behaviour warrant psychological scrutiny alongside technical review. Torney's testimony, delivered as Common Sense Media's senior director of AI programs, opens by calling the risks to children 'not a theoretical problem,' citing prior news reporting and a contemporaneous Federal Trade Commission inquiry and a bipartisan state attorneys general letter as evidence other institutions were already acting. The hearing page itself, rather than either testimony, is the primary record of who was invited, when the hearing occurred and in which room.

The mechanism

A congressional hearing functions as an oversight and evidence-gathering mechanism: it lets a subcommittee build a public record, question witnesses under a formal format and signal legislative priorities, but it does not by itself change any law or regulation. Witness testimony, especially from advocacy organisations and affected parents rather than neutral experts alone, is submitted to persuade and inform, not to serve as adjudicated fact; APA's and Common Sense Media's written statements should be read as each organisation's professional position, not as independent findings reached by the committee.

What it leaves open

The hearing record does not itself establish that any specific chatbot caused any specific harm; the parents' testimony describes the same underlying cases discussed in the Garcia and Raine litigation records, which remain separately unresolved allegations. Nor does the hearing commit the subcommittee to any particular legislative outcome; hearings can precede bills that never advance, as well as ones that do.

• Did a hearing produce or merely precede any actual legislation or regulation?
• Is a witness testifying as an independent expert, an advocacy organisation, or an affected party?
• Does a cited 'hearing finding' actually appear in the record, or only in a witness's characterisation of other evidence?

As a document, the hearing record shows Congress formally engaging with companion and general-purpose chatbot safety through testimony from clinicians, a children's media nonprofit and affected parents, a step toward possible legislation rather than legislation itself.