Considerate systems: the next big thing in artificial intelligence (AI)?

Has it ever struck you how inconsiderate your computer is? They flash up emails, while we're busy writing a report. They've no awareness of what we're doing and they disturb us.

Be more helpful?

The answer to this daily disruption could be a 'disruption manager' - intelligent software that manages our incoming messages to make us more productive, says Carnegie Mellon professor Ted Selker.

Selker's research in artificial intelligence looks at whether computers can become context-aware and more considerate (aka 'considerate systems'). Can they react to who's in the room and what's going on?

In the world of work, Selker's been experimenting with teleconferencing. Have you ever experienced the ups and downs of a teleconference? One person dominates and you're itching to chip in but you don't get a chance.

Shaddup you face!

Selker's been using intelligent agents to measure when one person has been talking for too long and others have been too quiet. When they judge that someone's dominating, they pipe up: 'Turn taking?' or 'Any thoughts, Suzanne?'

This automated system has apparently been successful at nudging people's behaviour. With the intelligent agents on the line, Selker found dominant people talked 20% less and quiet people talked 20% more.

Turn it down!

Selker's future projects may include a context-aware TV that senses how many people are in the room and whether they're talking. If 2 people start a conversation, the TV turns its volume down to a level that doesn't interfere. And if the TV senses your baby crying in the next room, it can turn the volume up or down as you wish.

And imagine an intelligent email program. You've written a rant and intelligent agents pick up on the language you've used. You get a nudge - a big red, ranty face on your screen - to reflect on what you're writing.

Ted Selker's seminar on considerate systems was at Queen Mary University of London, 12 November 2014.