The honest line in the dementia-robot pitch

University labs are quietly doing the unglamorous work the demo reels skip. Texas A&M is training a "digital human" to read apathy and biometric drift as early dementia signals, before a cognitive test would catch them. The University of New Hampshire is putting socially assistive robots inside actual homes, not lab mockups, and finding what every caregiver already knows: the messy, unstructured house is the hard part. The most honest line in the piece isn't a breakthrough claim. It's the researchers' own guardrail, that these robots "cannot and must not replace the human caregiver." That is the right altitude. The value here isn't a machine that takes over. It's earlier warning and a lighter load, so a family gets more good years at home before the harder decisions arrive.