Does AI Have a Future?

Does AI Have a Future?

By Christina Frederick-Recascino,
Associate Professor of Human Factors and Systems

In the recent Steven Spielberg film AI (Artificial Intelligence), a robotics company has developed the ultimate in artificial intelligence technology - the ability to program attachment and emotion into a robotic child that could be used as a replacement for biological children. To test the viability and marketability of their new product, the company gives a little mechanical boy named David to two parents grieving the seemingly permanent loss of their own son. The implications of introducing David into this family are enormous and address serious issues and complications that our society may face in the future if we look for personal AI applications.

Can we create an emotional machine? Attempts are certainly underway to do so, most notably at the Affective Computing Lab at MIT. However, current research into emotion and artificial intelligence is rudimentary and lacks many of the qualities needed for the full range of human emotional experience. Humans know how to recognize the verbal and non-verbal cues of emotion, use them to process information based on past experience and identify the correct emotion, gauge their responses to match or complement the emotional state perceived, and, based on further interaction, change an emotional response in seconds or minutes. Humans also are able to mask emotions when the need arises and feel several emotions as a result of a single event or interaction. As if these behaviors weren't hard enough to model, psychologists have yet to determine exactly how many human emotions exist. At best, AI researchers have used a physiological model to identify and project emotion and have typically limited the range of identified emotion to good-bad dimensions.

Should we even try to create an emotional machine? This is a moral dilemma not unlike those associated with recent technological developments in human cloning and stem cell research. Not only is there no easy answer to this question, it gives rise to further inquiries. Is emotion the sole domain of humans and other higher-order mammals? If a machine has emotion, does it possess humanity? Is it possible for us to become emotionally attached to a feeling machine? What are the repercussions of shutting down an emotional machine?

Before we go too far down the path to creating such a machine, it is imperative we try to resolve these moral dilemmas. For example, psychologists have already determined that spending too much time interacting with computers via Internet games, chat rooms, and e-mail creates problems in interpersonal relationships. Encouraging people to become emotionally attached to machines would perpetuate the social problem of our isolation and alienation from each other.

Maybe trying to create an emotional machine isn't what we should be focusing on. We humans have certain qualities that no machine has yet come close to emulating. These include creativity, decision-making in uncertain situations, crisis management skills, language, and perhaps emotion. Shouldn't we be trying to create technology that serves humans, supports our weaknesses, bolsters our strengths, and helps us create a safer, better society? An emotional machine is not needed to do these things, but better designed technology would certainly move us in the right direction.

For example, researchers at Embry-Riddle and elsewhere are working to create an automated air traffic control system that manages memory tasks that require an inordinate amount of cognitive effort. Another example of a useful artificial intelligence application is the creation of computer-based animals or cartoons for use in planning, scheduling, or searching large databases. Once they receive their parameters, these "social agents" perform their tasks in the background while the human user moves on to other activities.

As AI showed us David's life being played out in a human world, we were given a small glimpse of the repercussions of creating an AI child. The problems that occurred were amazing, and they took place in a simulated world created for us by Hollywood. But the real world is much more complex and uncertain than anything Hollywood can ever dream up. Do we really want to go there? Do we need to? Are the benefits of such exploration worth the costs? I don't think so.

The views expressed in "Perspectives" are those of the writer and not necessarily those of Embry-Riddle.