The Computer Tutor.

When I started writing this post, I intended to reiterate the points that I made in my first post about the relevance and applicability of Piaget’s ideas and principles, even in the face of new research. His idea of developmental stages and his laying down of what the majority of children can and cannot do at each stage has made it possible for programs like Sesame Street to be as effective as it is in educating the masses. This utilization of the different developmental stages is especially important because as pointed out in the article by Lepper et al., the luxury of having a personal tutor is limited to a select few.

I was going on to talk about how I thought that although it is possible that technology may one day come close to being as good a tutor as a human being, I doubted that a machine could ever be equally effective. As Lepper et al. said, the computer might be able to meet the cognitive requirements for tutoring, but the socioemotional aspect to tutoring seems to be a bit more out of reach. I was then listing a few reasons why I thought the socioemotional aspect would unattainable for computers, and they all revolved around the computer not having Theory of Mind. The more reasons I listed, the more plausible the idea of computers having Theory of Mind actually seemed!

This change of heart occurred when I wanted to say that teaching our young was a biological characteristic, and was a task that could not be taken over by an inanimate object. It struck me then that this was a contradiction to what I had said I my last post about psychological and biological mechanisms being distinct. That being the case, it is possible for a mechanical object to be a mentalistic agent, perceive the mental states of other mentalistic agents, and to react contingently even though it is not a live being.

One of the main barriers to a child’s development of Theory of Mind is the level of inhibitory control needed. Theory of Mind has been found to increase when a child’s inhibitory control increases (Carlson & Moses, 2001). This is because the child needs to repress the most salient beliefs and desires available to them – that of their own – and identify the beliefs and desires of others. Computers would not face these problems because for one, they do not have beliefs and desires of their own per se. Even if they did, computers are machines – they should not have any trouble inhibiting their own “mental states” and prioritizing the input they are receiving about their tutees.

Beliefs and desires are sometimes presented inaccurately on purpose. Computers can be programmed to be especially sensitive to social cues such as facial expressions, body language, and even physiological states. All this information combined might give the computer an even better representation of the tutee’s emotional and mental states than would be available to even the most expert human tutors. Thus, modulating a problem’s subjective and objective difficulties “intuitively” would be possible, and maybe even easier.

It is true that with a human tutor, modifications can be made to teaching styles according to the different personalities of different tutees, and this flexibility might be a problem for computers. Something similar might be accomplished, however, by including a personality inventory with the teaching software, and different teaching styles modeled on the styles used by expert tutors can be utilized appropriately. As for creating playful competition, the fact that many computer games are played against computers, and that children are motivated to do so, show that this aspect of human tutoring is achievable in computers too.

Given the rate of technological advancement today, it is possible that children all over the world will be given the opportunity to have one-on-one tutoring sessions with computers, tailor-made to suit their individual needs, relatively soon. When such a computer or computer software is created, it would be interesting to compare these computer tutors to human tutors to see if the biological feature of human tutors plays an imperative role in expert tutoring, or if psychological mechanisms are sufficient.