II
Cogs
Emotional-learning frameworks for humanoid robotics — so that as machines move into homes and clinics, the interaction is safe, empathetic, consent-bound, and human-led.
Cogs studies the affective layer of humanoid robotics — the signals a machine should attend to, the responses it should withhold, and the disclosures it owes the people it interacts with.
The program treats empathy not as a feature to be added but as a system property to be designed for. That means measurable behavioral outcomes, third-party audit of decision policies, and explicit consent flows for any deployment that involves vulnerable populations.
Part of that agenda is assistive: Cogs studies affect-aware tools built with — not for — people the field tends to overlook, including autistic adults navigating sensory load, daily living, and communication on their own terms. The work is held to the robotics’ own standard — it expands a person’s autonomy, and never quietly manages the person who depends on it.
Cogs is conducted in partnership with academic robotics labs, a small number of clinical research sites, and self-advocacy organizations. No deployment touching a vulnerable population is considered until a peer-reviewed safety case has been published.
What the program is asking.
01
What does it mean for a humanoid robot to be honest about its own uncertainty?
02
When should an affect-aware system defer to a human, and how is that boundary engineered?
03
How do we build assistive technology that supports autonomy rather than quietly managing the person who depends on it?
The other programs.
Proposals welcome.
The Foundation accepts grant proposals, partnership inquiries, and fellowship applications from researchers whose work intersects with this program.