Enter the maze

Future friends who get around

a woman in a montage of different emotions

We can all do with a little help at times. Relying on someone else for their companionship is a fairly basic human need, but perhaps in the future our companions will be artificial. A large European Union-funded research project called LIREC (that stands for Living with Robots and Interactive Companions) is setting out to discover how we can create a new computer technology that helps us form long-term relationships between humans and synthetic companions in real social settings.

Out of the laboratory

LIREC’s artificial companions won’t be studied in a laboratory though. They’ll be going into homes and offices as the new generation of artificial companions, in the form of robots or as graphical characters on screen. They will be socially aware, which means they will have some idea of how to act properly in particular situations, like any good and trusted companion would. The companions will be tested in three main environments: a real house where real people live with helper robots, an office in a university where robots assist in the day-to-day work, and as playmates for kids, playing games like chess. The researchers in LIREC hope the people will start to form relationships with the robots. They may start to see them as a useful part of the world around them – not able to do everything a human can, but still a valuable part of their social circle.

Important memories

At the heart of these companions will be state-of-the-art artificial intelligence, but this needs to be socially smart software. Imagine your artificial companion knows what you did last night. That’s fine – your pals know that too. You have to share some information to build up a bond with someone. Now suppose a stranger (or your parents) ask the companion what you’ve been up to. You would expect that a friend would be careful about exactly what they would reveal in each case. That’s called tact: knowing what to say, when and to whom. We humans do it all the time.

This ability to thoughtfully filter information is part of the social glue that keeps us together. The artificial companions need to be able to do this too, so the software needs to be able to understand what is appropriate in different situations, and act accordingly. It’s a very challenging problem. Have a go at listing how you would decide who should know what about a range of different topics. It’s tough. One way to help design the software is to look at the way human memory works, how information gets connected and modified, and whether you remember or forget a particular incident. At the core of it, it seems that people don’t work like a search engine.

We can’t instantly recall everything that’s happened to us, but that leads to a fascinating question. Would we want total recall from an artificial companion’s memory, or is forgetting stuff what we would expect?

a cyborg extends its hand out of a computer screen

With friends like these?

LIREC is also about understanding how these artificial companions can fit into the real world, and what they should and shouldn’t do. Until now artificial companions have had fairly limited abilities to create long-term, meaningful social interactions with humans in real social settings. They have tended to suffer from the novelty effect; they are fun for a bit then get boring. But as LIREC sets out to create artificial companions that can hold a long-term relationship in people’s lives, the researchers also need to consider issues like personal privacy and ethics.

What safeguards should be built into the companions? Are there no-go areas? This research is as important as creating the software and hardware.

Moving on, getting mobile

Today’s robots are clever but still limited in what they can do, so LIREC will push things one stage further. LIREC’s artificial companions will migrate. The same companion that’s in the robot body at one moment will be able to move into a mobile device, like a phone, at another. The software will be able to adapt to its new environment, but hopefully you will still be able to recognise it’s the same ’individual’ you knew in the robot body. When you’re talking to a friend on the phone you can’t see their face or body language, but you still know it’s them, and you still feel the same attachment to them. This is cutting edge stuff. What does it mean to be the same ‘person’ in a robot shell or on a computer screen?

How does the software adapt when it’s being used on devices with different electronic memory capacity? What’s essential to transfer and what can be left behind? What can be learned and expressed in one body and moved back to the other? Finally, and maybe biggest of all, what makes a long-term relationship work? It’s not just about technology, LIREC is also asking deep questions about human relationships.

The future

You have probably heard the term ‘user friendly’ when applied to technology. That means a technology is developed focusing on the needs of the human user, and advances in this field have changed the face of consumer electronics. The ease of using an iPod has been one of its main selling points. LIREC hopes to go a stage further and develop ‘artificial companionship’. The technology will allow future synthetic products to build and maintain a long term, meaningful link with the user; from personal digital assistants to home care robotics, from intelligent offices to entertainment. Understanding the essential nature of relationships and how this can be incorporated in intelligent software could change future applications as much as the adoption of user-friendly technology already has.