In my first discussion of "nanny-bots", or robot caregivers for children, I referred to the ethical concerns noted by Dr. Noel Sharkey in his article, "The Ethical Frontiers of Robotics" (Science, Vol. 234, 19 December, 2008, pp. 1800-1801). I agreed with Dr. Sharkey's worries, especially with respect to infant care. Everything we know about infant and toddler development suggests that robots could not provide the sensitive, responsive social communication that babies need and that human caregivers typically give.
On the other hand, it's an interesting thought that robots might not only be helpful, but even do a better job of some child-care tasks than humans do. For example, groups of adults do not necessarily do a good job of supervising children at play, because the adults become interested in each other, begin to talk, and may forget to watch the children as each thinks the others are paying attention. Several robots "watching" from different angles could do a much better job of noticing a hurt child or the approach of an intruder.
Specialized robots could deal well with situations in which adults may be too hurried or distracted to do a good job. For example, how about a baby bath robot? Such a device could sense appropriate water temperature and prevent burns to infants' sensitive skin. For babies who can sit firmly, a bath-bot could even monitor and safeguard the baby's position and breathing while the caregiver left the room, which would otherwise be a dangerous action. A temperature-sensing robot could also provide continuing information about a child's fever and about the temperature of food or milk.
What about a stair-top robot to replace baby gates? The gates currently in use are all too easy to leave open as someone rushes up or down with the intention of returning at once. A stair-bot could stand aside as adults, older children, and pets moved freely up and down, but reliably block the movement of a crawler or toddler who was heading for a fall.
Although it seems unlikely that robots could ever protect infants so well that houses would not need to be "baby-proofed", nevertheless they could play important roles in emergencies by limiting a baby's movement. How do caregivers now handle the situation if a toddler is moving around the house, when crying outdoors alerts the adult that an older child has fallen off his bicycle? They may take the toddler out with them, a very awkward solution, or they may have a playpen or crib handy and dump the little one in it before going to administer first aid. An emerg-o-bot could be programmed to keep the toddler in a limited safe area, and turned on as the adult rushed to the emergency. (Family day care providers, who may be the only adults caring for groups of 4 or 5 young children, might find such a robot especially useful.)
A robot that could accompany an older child as she went to a friend's house or played outdoors could signal the child's position and call for help under certain circumstances. What shall we call this, a Lassie-bot?
By the teen years, most robot-guarded individuals would be likely to "lose" or sabotage their robot caregivers. Nevertheless, robot care could be built into some situations. For example, a car-bot-without which a car could not be started-could assess how many passengers a teen-age driver was carrying, whether seat belts were fastened, and even whether the driver's breath indicated alcohol consumption.
All these nanny-bot uses are directed to children's physical safety. Real developmental concerns, such as speech experience as a foundation for language, or emotional interactions to form attachment, remain the job of human caregivers for now, and possibly forever.