FOR IMMEDIATE RELEASE

Oct. 15, 2026
Caregiver Action Network
Contact: Allison Borthwick
Email: aborthwick@caregiveraction.org
Phone: 202-780-9978

————————————–

WASHINGTON, DC — Caregiver Action Network (CAN) today launched The Humanity Baseline: CAN’s Framework for Responsible AI in Caregiving, a public statement of how CAN will engage with AI and support family caregivers navigating it. The framework explains how CAN will approach AI-enabled tools and the questions caregivers can ask about them, while keeping human judgment, choice and the dignity of people receiving care at the center.

Caregivers are already testing how AI might help them find information and manage daily responsibilities. In CAN’s 2026 Caregiver Tech Insights Survey, 59% of 272 family caregiver respondents said they had tried or regularly used AI for caregiving tasks. Trust, privacy and cost remain important barriers for those considering new tools.

“We’re not asking caregivers to choose between useful technology and the human support they depend on,” said Marvell Adams Jr., CEO of Caregiver Action Network. “AI can take some of the information overload off a caregiver’s plate. It cannot provide human compassion or replace a trusted relationship; consequential decisions still need meaningful human oversight. This framework puts a clear stake in the ground for how CAN will approach AI and the questions we’ll help caregivers ask as they use it.”

The Humanity Baseline outlines five principles: humanity and choice; trusted, reliable and verifiable information; privacy, transparency and control; co-design and continuous learning; and evidence and accountability in real caregiving situations. It addresses both tools caregivers use directly and AI embedded in services. The framework is CAN’s living guidance—not a product certification, compliance standard or endorsement of any particular tool.

Ali Madad, Founder of GiveCare, creator of InvisibleBench and an advisor to CAN’s Caregiver Innovation Council (CIC), helped shape the framework’s attention to real-world risks and the caregiving relationship.

“Care is a relationship, not just a task to automate,” said Madad. “An AI tool can sound helpful while missing distress, crossing into advice it should not give or overlooking the person receiving care. I was glad to help CAN shape a framework that asks what happens in real caregiving situations, not only whether a tool performs well in a demonstration.”

Adams moderated a May 19 discussion with Madad and other leaders at the Elizabeth Dole Foundation’s 11th Annual National Convening on “The Future of Caregiving: Role of Technology, AI, and Innovation.” The conversation explored practical uses of technology, safety, trust and the need for caregiver input. Watch the panel conversation.

The launch comes ahead of the free, in-person AI Skills Jams for Caregivers & Community Health on November 10. Hosted by OpenAI Academy in collaboration with CAN, the Alzheimer’s Association and the American Diabetes Association, the workshops will offer hands-on, nonclinical practice with ChatGPT, including resource navigation and caregiver coordination—not diagnosis or treatment advice. CAN hopes to use what it learns from caregivers and future training opportunities to inform more practical, plain-language guidance over time. Register for the Oakland, Tampa, Pittsburgh, Atlanta or Detroit Jams.

Read The Humanity Baseline: CAN’s Framework for Responsible AI in Caregiving