Deploying social robots in care settings requires ethics review beyond standard usability testing

What it is

Deploying a social robot in a care setting — for people with cognitive impairment, frail older adults, and similar populations — requires ethics review beyond standard usability testing. Even without performing any medical act, the robot can influence consequential care decisions through monitoring, advice, and the affective relationship it builds, which is the fundamental difference from an ordinary consumer product's ethical procedure.

Why it happens

The need for additional review comes from diminished capacity to give informed consent. The core premise of an ethics review process is that the person being served can understand and freely consent to the risks involved, and people with cognitive impairment or significant cognitive decline often lack full capacity to do so. In that situation, risks around attachment, privacy, and continuous monitoring — a camera-equipped companion robot, for instance — cannot be waived simply because "the user agreed"; they require additional mechanisms such as proxy consent from a guardian and review by an institutional ethics committee. This is a structural difference in ethical procedure between care and ordinary consumer settings, not a matter of being somewhat more careful.

Where it stops holding

How strict the review needs to be scales with the degree of cognitive impairment and the type of risk involved — whether physiological or location data is collected, whether the product affects medication or care decisions. Not every product aimed at older adults needs institutional-level ethics review; it is triggered only when informed-consent capacity is compromised or the product affects care decisions. Passing review also does not guarantee the product will cause no harm in actual deployment — review is a threshold to clear before deployment, not an endpoint.

Applying it

Social robot projects aimed at people with cognitive impairment or significant disability need a guardian or family informed-consent process built in at the design stage, with explicit rules on who has authority to consent on the user's behalf. Features involving data collection — audio, video, location — need a separate privacy impact assessment rather than being folded into general usability testing. The concrete check is submitting the plan to an institutional ethics committee before formal deployment and adjusting the scope of data collection and the consent process based on its feedback, rather than treating review as paperwork to complete after deployment is already underway.

Related

• Same group: X6.05.1 Children and older adults form different mental models of what a robot is and can do · X6.05.2 Relying on a social robot for companionship can displace rather than supplement real human contact
• Nearby: X6.03 Emotional expression · X5.06 Presence of vulnerable populations
• Search terms: informed consent · institutional review board · socially assistive robotics · human-robot interaction