AI Tools for Infant Airways and Dementia Care

Research projects at five US universities are testing Meta's AI and headsets for infant tracheostomy monitoring, dementia communication and heart failure risk.

Updated on Nov. 3, 2026 in Alzheimer’s

Isometric editorial illustration of a matte white wearable headset with blank lenses beside a thin pale tube on a plain grey surface. — Meta has partnered with five U.S. universities on research testing its AI and mixed reality headsets for infant airway monitoring, dementia communication, and heart failure risk. AI Illustration. Upload story photo >

Live Poll

Do you trust big tech companies to help develop safe medical technology?

Meta has partnered with five US universities on medical and scientific research projects that will use its AI models, smart glasses, and mixed reality headsets. Projects include a system to detect tracheostomy tube displacement in infants, conversational AI for dementia care, and a heart failure risk model.

Why it matters

The projects target conditions where timely detection or better communication could change care, including life-threatening airway emergencies in infants and communication breakdowns in people with dementia. Using open-weight models lets academic teams build and test tools on their own data.

The Eyes-On project has published a proof-of-concept, but the source reports no performance data. The other two projects are described only as in development, without published results.

The players

The details

The Eyes-On system is a computer vision tool designed to spot when a tracheostomy tube in an infant has shifted out of position, and it is intended to work with bedside cameras and edge computing. The Ohio State heart failure project is designed to predict HFpEF risk from abdominal CT scans, medical records, and ECG data, and the team is building it to handle missing data. The dementia project focuses on conversational AI that can better accommodate the communication difficulties of people with dementia.

Announced November 3, 2026, with work on the projects described as underway.
Meta says it will share more as the projects get underway.

Health Landscape

Academic groups are increasingly building clinical tools on open-weight AI models rather than proprietary systems, which may speed research but leaves validation as the main hurdle. Early-stage monitoring and prediction tools like these must still show accuracy in real clinical settings before they can change standard care.

For now these are research projects, so no change to screening or care is expected yet. Families of infants with tracheostomies or adults with dementia may eventually see these tools, and it is worth asking a clinician what monitoring or communication support is available today.

The takeaway

Artificial intelligence is moving into specific clinical problems, from infant airway safety to dementia communication, but these tools are still in early development. Readers should expect validated results before any of them reach routine care.