Kings County approves AI‑assisted fidelity monitoring for child‑welfare motivational interviewing amid privacy and workforce questions

Kings County supervisors voted to approve an agreement with LISN (branded in staff materials as Listen) to provide AI‑assisted fidelity monitoring for motivational interviewing used by county child‑welfare staff.

HSA program manager Natasha Endress told the board the county chose motivational interviewing as its evidence‑based model for prevention services required under the Families First Prevention Services Act. She said the county must monitor fidelity using established, research‑based tools — the Motivational Interviewing Competency Assessment (MICA) and the Motivational Interviewing Treatment Integrity (MITI) — to claim federal reimbursement for prevention services beginning in 2026.

“LISN is a secure artificial intelligence or AI assisted platform that uses 2 evidence based tools to measure motivational interviewing fidelity,” Endress said. She told the board the tool scores role‑play sessions created for training, stores recordings on HIPAA‑compliant servers in the United States, allows deletion of recordings after scoring, and that only de‑identified scores would be retained for fidelity tracking.

County staff said the first year of the contract will use Families First Transition Act funds that must be spent by Sept. 30, and the second year will use Families First Prevention Services Act funding distributed by the California Department of Social Services. Staff also said a sole‑source procurement was justified and approved by the purchasing chief information officer in February because LISN was the only option offering automated fidelity scoring at scale.

Board discussion and public comment

Supervisors asked whether the platform would be used to record actual client interviews; Endress and HSA staff were explicit that the program would be used only for scripted role plays and not for recording clients. Endress said, “we are not even gonna say any kind of client information. It's gonna be more so just to allow the social worker to practice their skills.” In response to privacy concerns, staff confirmed that data are stored on HIPAA‑compliant U.S. servers and that role‑play recordings may be deleted after scoring while numerical fidelity scores would be retained.

Several supervisors expressed caution about wider AI use and workforce impacts during the discussion; one supervisor told the board, in effect, to track what the technology returns and not to assume it is infallible. Lisa Lewis of Kings County Behavioral Health spoke in support of tools that reduce the cost and time of fidelity monitoring, saying training and expert supervision are expensive.

The board approved the agreement by roll call; four supervisors voted in favor, one was absent.

Why it matters: Monitoring fidelity is a federal requirement tied to the Families First funding stream; the county’s ability to claim Title IV‑E reimbursement for prevention services depends on documenting use of an evidence‑based model and monitoring fidelity. County staff said the platform will provide faster, repeatable feedback so supervisors can focus training where it is needed.