Work in teams of three or four people based on the number in the upper right hand corner of this page.
Consider these things that a computer system of the future might do:
Use this list to guide your work on Tasks 1-4.
How should this system interact with human clients (users)?
Would you be comfortable using such a system? Why (not)?
Be sure to give the *why* behind each of your answers!
Submit your group's answers to the discussion questions.
We will clarify any questions that you run into during the exercise.
Building computer systems capable of doing these kinds of tasks is one of the goals of artificial intelligence, an area in computer science that interacts with disciplines such as cognitive psycology, linguistics, philosophy, and engineering. Our goal here is not to study AI in depth but rather to consider it from the perspective of a society in which AI systems might play a role. What are the implications for people? What are the implications for society as a whole?
A couple of years ago, AI of a sort reached the mass media when IBM's Deep Blue, a special-purpose chess playing computer, played two matches with world chess champion Garry Kasparov. Deep Blue won the second match and created a firestorm of questions about whether humans had been surpassed in the intellectual arena by a machine. What do you think?
Defining AI is difficult, because we aren't sure what we mean by "intelligence" when we talk about humans, let alone machines. Marvin Minsky, one of the founders of AI, defines AI as "making computers do what humans are good at". Others say, "programming a computer to do something that, if done by a human, we would consider intelligent". (Notice the subtle difference in these definitions.) Finally, the most tongue-in-cheek definition I've ever seen is, "if we can do it, then it's not AI". Kind of tough to make progress as a researcher in a world with such a mindset.
Here is my favorite "technical" definition: AI is the study of the computations that make it possible to perceive, reason, and act in a complex environment.
How does AI differ from psychology? Philosophy? The rest of CS? It shares with all three questions of interest and methods for seeking answers to these questions. But it is different from each.
What kinds of problems does AI study? Given the breadth of definitions, you should not be surprised to find that the problems we study come from a wide spectrum. The items in the exercise today give a pretty good sampling. Studying formal systems such as mathematics and game playing was more dominant early on in AI, but now researchers tend to focus on more realistic tasks. (In many ways the formal tasks are too easy--because they are formally defined!)
What does an AI scientist do?
The key question I ask you to think about for this course is: How would "intelligent" computer systems affect the way we interact with the world? With each other?
Keep in mind, too, that what AI seeks to do may seem magical today, but today's "common" software would have looked like to magic to people in 1955, 1965, 1975, 1985--maybe even 1995!
In our group we discussed the questions and thought that the easiest task for a computer system would be the board games, since many computers already have them. Like chess. When we got into our large class discussion the class agreed with our group on board games being the easiest.
Our group thought the hardest would be playing with children or recognizing the human face. It is hard to interact with children and having a computer do it would probably take away from the parents responsibility. I also think it would create mental or behavioral problem with children. I thought that recognizing the human face would be difficult too because we are always changing, from year to year a lot of people change. In the large class, playing with children was discussed as the hardest thing to do people mentioned that it would be hard to interact.
Our group thought that recognize and generate spoken language would be the most beneficial it would cut down on the work that people had to do when using a computer, and the least beneficial was the board games, sure it is fun to play games but it doesn't really help anyone. The class thought the most beneficial thing was diagnosing a human illness--it could be more accurate. Like the computer (Mycin) and how it could diagnose a blood disease more accurately. I personally wouldn't want to be diagnosed by a computer, because with a real human doctor there is more interaction and they often respond with questions to really understand what is wrong with you.
The five things that we selected were: recognize faces--how would the computer get the image in the first place, and would we have to keep updating it, university teaching schedule--would the computer get input from the faculty and students, solving calculus problems-- students may not know how to actually do the problem and let the computer do all the work, legal argument--would it take in the facts of the case or would it be standard argument/ if this is the problem then this is the argument, and play with children--is it going to take the place of the parents and the parent/child time together.
Our answers to these were: 1. give facial info and scan a picture of yourself every couple of years. 2. teachers and students give input, and establish their needs. Do the schedules based on needs. 3. require students to show hadwritten work. 4. take into account all facts and information from each case. 5. Time limit of computers--but it is really up to the parents to spend time with their children.
Question number five was answered by my group as the client would type in the problems and the computer would output a solution. We said we wouldn't be comfortable with the computer because the doc. aske questions to get the full understanding of the problem that we could have forgotten to type in and we might have gotten a wrong solution then.
And we said we would not want a computer playing with our children because it is the parents job and computers may not have the sense to "raise" our children the way we want them to. We would also not to be diagnosed by a computer because it is dealing with human life and if we don't give enough or the right information it could potentially hurt us.
The hardest for the computer system to implement to me was kind of a culmination of the choices. During the class discussion, we talked about the computer playing with kids, which is different from kids playing on the computer. We talked about how kids need nurturing, have revolving emotions, the ability to be mobile, and social interaction. I think it would be difficult to create a system that could do all the above things and recognize the child's specific needs. Dr. Wallingford mentioned the making of sociable robots, and how these robots are not people but pets. The people are looking into the needs of the robot and the needs of a human.
The second question to consider was which option would be most beneficial to humankind. I thought right away that diagnosing human illnesses would be the most beneficial. However, it would be tricky, because people can leave out symptoms and we also lose the interaction and the familiarity of family physicians. I think people feel better being diagnosed by a human, rather than a computer. For one, they have someone to blame if the diagnosis is wrong, and two, they have someone to sit and talk in layman's terms about their diagnosis.
This leads to the final question, Is there anything that we should not let computers do? I think mostly everyone in class felt that diagnosing human illness should not be done by computers. There were a couple of people, including Dr. Wallingford that would agree to allow computer systems to diagnose human illness. His argument was that the possibility that the computer would be more right on about the diagnosis was enough to let the computer diagnose his illness. I thought after his argument that maybe it would be okay to let computers doing the diagnosing. However, I am still very much for personal interaction with a human doctor, and maybe the computer can be a help to the doctor and make the diagnosis much quicker.
The list of things we were asked to consider a computer doing in the future included: [see above].
Our group decided that from this list, the easiest to implement would be playing board games because this already exists and doesn't have too many social implications that we could think of. However, we thought playing with children would be the hardest, as did most of the class. This would be difficult because it would require computers to deal with human behaviors and emotions. Children are constantly bouncing from one place to another and it would be difficult for a computer to keep up and provide the nurturing and guidance children require. I don't think computers should be allowed to do this because there has to be a point when we stop looking to technology to make our lives easier. We should take on the responsibility of caring for kids and be there to guide and support them, rather than pawning them off on a robot.
We felt the function most beneficial to humankind would be recognizing spoken language simply because it reduces the human workload without any personal implications. It doesn't take away from any valuable personal interaction and could make our lives a little easier by simplifying repetitive tasks.
An interesting topic that we spent quite a bit of time discussing was whether or not computers should be allowed to diagnose human illness. A couple people in the classroom suggested they'd feel comfortable with this process; however, the majority would not be comfortable. I am with the majority and think this is a scary idea. I don't really know how these systems would interact with people, but one way our group mentioned was to have the patient input the symptoms they have and then have the computer generate a response. We didn't feel this would be successful because you may only input a limited number of symptoms and disregard some important factors for the diagnosis. If you were at a doctor, they could ask you a series of questions in order to get the full story. It's just a scary idea to think a computer may be making life or death decisions. But, I guess 50 years ago people never thought we'd be where we are today so it is all a part of change.
After thinking a little more about the topic of computers diagnosing illness, I came to realize that maybe it would be possible for the computer to act like a doctor, asking more questions to get the full picture also. This would be more comforting than listing 2 symptoms and receiving a diagnosis. Another point that was addressed is that humans are not perfect either and there is no way they can possibly hold all the relevant information in their heads. It's nice to be treated personally, but if computers could help improve health in heavily populated areas, maybe this could be beneficial.
In conclusion, there are endless possibilities for what computers may do in the future. This could be beneficial to a certain extent; however, I think we need to take into serious consideration the social, legal, and political effects. We already live in a world where we just send an email to the person sitting next to us rather than getting up to talk to them. I strongly believe we need to make sure we continue to develop the personal relationships that are a part of everyday life. Technology is nice for repetitive, meaningless tasks, but I think we need to continue to have personal interaction for many daily activities.