Making homes smart

Genevieve Bell, Senior Researcher, Intel Research
Genevieve Bell, anthropologist and top researcher at Intel, was star turn at a fascinating seminar run by the iSociety research group at the Work Foundation. The topic was ‘the smart home’. Bell’s current project is aimed at understanding the use of technology in homes across Asia. She understands how deep-rooted religious, social and cultural practices influence the reception of technology in homes.

If there was any conclusion it was that attempts to add intelligence to our homes had better be respectful of the needs and beliefs of their inhabitants, otherwise they’ll just join the long line of forgotten household gadgets – egg slicers, the thing with the plunger that makes cream. The whole thing reminded me of Stewart Brand’s How Buildings Learn, a book about the way humans adapt their habitations and continually turn them to new uses, usually despite the architects and planners. If we’re to augment our homes in the ways advanced by the technology vendors the kit will need to fit properly into the lifecycle of a home.

It should be cheap, reliable, modular, interoperable and easily replaced or upgraded. Homes exist on timescales measured in centuries, a family will occupy a home for decades and make continual, incremental changes. Wi-fi hubs, entertainment systems and central heating controllers can’t aspire to this kind of longevity but should aim to fit into these cycles unobtrusively, helpfully, without expecting families to bend to new ways of living.

Houses will get smarter in much the same way they acquired their current intelligence – looms of copper wire, electric light, telecoms, wireless and television, modern additions like insulation and heating, older ones like plumbing and drainage – piecemeal and over a long period of time. The best the smart home business can hope for is to get their kit onto the makeover shopping list, to become must-have lifestyle items and to slot into the cycle of home fashion… like decking and rag rolling, really.