Undeterred by the persistent worries of privacy advocates and government officials that it knows too much, Google hungers for more data. To augment the information the company collects from its users online -- the links they click, the searches they make, and related metrics -- Google's researchers are looking beyond the Internet.

A recent paper co-authored by Google researcher Bill N. Schilit, and computer scientists Jeonghwa Yang (Georgia Tech) and David W. McDonald (University of Washington) proposes "home activity recognition," or tracking people's activities at home through home network interactions.

"Activity recognition is a key feature of many ubiquitous computing applications ranging from office worker tracking to home health care," the paper explains. "In general, activity recognition systems unobtrusively observe the behavior of people and characteristics of their environments, and, when necessary, take actions in response -- ideally with little explicit user direction."

The goal of such monitoring might be to "remind users to perform missed activities or complete actions (like taking medicine), help them recall information, or encourage them to act more safely," the paper suggests.

As applied to the elderly, such monitoring might seem entirely sensible. Others might find such oversight Orwellian.

Is it comforting or frightening to think of Google looking after one's health? "Information about household activities can even be used to recommend changes in behavior -- for example, to reduce TV viewing and spend more time playing aerobic games on the Wii," the paper suggests.

Just wait for the pop-up menu that says, "Type faster, porky."

Whether the future Google is exploring is benevolent, malevolent or just the way things will be, such a scheme raises questions about sanctity of the data describing one's activities at home. How would that data be protected? Who would have access to it? What would prevent it from being subpoenaed or stolen?