ITU’s Sofia put on an impressive display at the AI for Global Good Conference but it raises an Important Question, Should we humanize AI?

Artificial intelligence. The concept has fascinated science fiction writers and scientists for decades. With some hailing advancements in AI as the dawn for a new era of man and others believing that an advanced AI could spell the end of our species. There is a lot out there about the benefits and potential pitfalls of AI but a video at the AI for Good conference of ITU’s Sofia raises the question; should we humanize AI?

Sofia is undoubtedly impressive, even if there are shades of uncanny valley. It is responding in real time, to random questions from the facebook audience. On top of that it can make recognizable facial expressions and with the addition of hair and some refinement it would look almost human. The question is why ITU is interested in making Sofia appear so human.

There have been a few attempts to create human-like robots and there are good reasons for doing so. Many believe that we can use these kinds of robots to plug the growing gaps in our social care systems, especially in light of increasingly aging populations in nations like the USA, Japan. The belief is that if we humanize AI and robots they will make better carers. After all, nobody wants to be taken care of by some terrifying metal contraption.

AI has already proven its use in a number of fields. Scientists have used it to help identify cancer and there have been trials at using a robot to replace a doctor. Dr Nao have proven that patients are willing to trust a robotic doctor but his role would likely be mostly to provide support for overburdened staff.

There are clear applications for robots and AIs in roles that we would traditionally consider human. Given that fact, why should we be concerned about humanizing AI? There are the traditional concerns for jobs and livelihoods but there are also deeper ethical concerns. Let’s take the example given to one user by ITU’s facebook feed, using robots like Sofia in caregiver roles.

The care gap is a huge problem in most Western societies. As populations age and the family unit increasingly fragments care for the elderly is going to become more expensive and more difficult. It is tempting to think that we can give them a human-like Robot like Sofia to look after them but it would bring its own problems with it. While Sofia and AI like her would look and appear human they wouldn’t be human.

By putting robots in the role of carer we are making them more human but we risk robotizing the vulnerable people in its care. In essence we are saying that they are another commodity to be managed, rather than individuals with specific needs.

AI excels at many things and it has already proven that it is an invaluable supporting tool for human scientists and professionals. It is capable of sifting through vast amounts of data at speeds humans could only dream of but I do not believe that we can view it as the panacea  for all ills. We need to remember that while we can humanize AI that doesn’t make it human and that doesn’t mean that it can replace contact with another human being.

AI and robots are invaluable when it comes to supporting us but do we want them to take over certain roles in our society? It is certainly true that robots can fulfill roles like caregivers, teachers and doctors but I think we need to stop and ask the question.

Does humanizing AI risk robotizing the people it is supposed to help?