Get Full Government Meeting Transcripts, Videos, & Alerts Forever!
Get email alerts on the Ai Design Trust topic
No spam. Unsubscribe anytime.
Caregivers, direct support professionals urge privacy, accuracy and personalization in AI tools
Summary
Caregivers and direct support professionals on the ACL webinar called for AI tools that save time, protect privacy, explain errors, and allow personalization and user control; recommended features ranged from earbud‑based on‑the‑job training to upfront data‑use disclosures and accountability when mistakes occur.
Caregivers and direct support professionals who joined the ACL webinar roundtable described concrete ways AI could help and listed conditions for trust.
Panelists said the most valuable AI features would be those that reduce repetitive administrative work and save minutes that compound across a caregiver’s week. Nicholas Smith, a direct support professional and NADSP board member, described small, practical examples: "Just imagine, like, for something as simple as, like a recipe... I want AI just say, hey listen, give me a simple recipe for a quick lasagna," he said, noting even small time savings can change a day. Kelley Shepherd, a direct support professional, proposed audio-based on‑demand training delivered via earbuds so staff can learn while working.
Panelists repeatedly raised privacy and accuracy concerns. Annelliese Barron, Executive Director of One Family Foundation and a family caregiver, said she fears personal information could be used in marketing without consent: "I'm always afraid that my name and my face is going to appear on some marketing material... without my consent." Angela Kouters, who worked in government affairs and cared for family members, said she would expect an immediate explanation and fix after any AI error: "If there's a screw up, I want a note explaining to me how you fixed it and how it's not going to happen again."
Speakers recommended design features to increase trust: clear, upfront explanations of how collected information will be used and separated or aggregated; customizable listening and sharing settings that let workers separate 'work' profiles from 'home' profiles; and embedded professional guardrails (for example, training content or ethical codes) to reduce mistakes and bias. Nicholas Smith suggested granular personalization so a worker can choose when the tool listens and when it does not.
The roundtable highlighted a tension: caregivers are open to AI when it demonstrably reduces time and complexity, but widespread adoption requires protections for privacy, transparent accuracy, easy onboarding and an obvious value proposition.
CAN and ACL closed the webinar by thanking participants and directing attendees to posted resources and the Slack workspace for follow‑up.
