Assignment XII
1.Turing…
The Imitation Game is far from being a good substitute for figuring out if machines can think because of the broad definition of the word ‘think’. Thinking, although intangible and does not require physical and voice representation, does require the ability to reason and understand, it is not merely giving answers to questions. Questions in an imitation game for computers are aimed at identifying if computers can think like men where reason and logic are required and it could be successfully used as a substitute in that regard. To imitate human conversation by following logic is not nearly as accurate as saying that the machine has intelligence or can think. The computer is given the logic by a human and so the reliability of using this game as a signal is lost there. If the signal of thinking and intelligence is restricted to questions and answers within the logic then yes, the game could be a reliable signal of intelligence, more like a reliable signal of whether the human is a good programmer and engineer! However, certainly we go by the definition that thinking and intelligence is way beyond just logical answers to consider the imitation game as a good signal of intelligence.
2.Weizenbaum…
The Rogerian or Person-centered therapy places the focus on the ‘client’ (not patient) to get him or her to talk about their ideas and feelings. Here, therapy and the role of the doctor is not to be prescriptive but let ting the patient talk about anything. Since Eliza is based on this theory, it does not require initiating a new chain of thought or generating ideas to give the client. ELIZA asks questions based on keywords in what the patient says and forming simple repetitive questions to hear the client’s point of view. Thus, people could communicate effectively with it knowing that it is substituting a therapist and so is not expected to have a normal social conversation with opinions, original thoughts, and retorts. This I think made people realize that ELIZA does not have intelligence of a human being because its responses are programmed.
I think that if a person does not know they are communicating with a chatbox, the conversation could be interesting and engaging but up to a certain point only, at which they may realize that the pattern of responses by the chatbot. Communicating with a chatbot whose model and structure is known, might have different results in a communication. The person communicating would know what responses/questions to expect and can direct the conversation the way he or she wants to get the desired outcome from the communication.
3.Donath…
Intelligence as defined by Webster is the ability to learn, understand, and reason all of which require mental ability. Because intelligence is intangible, communicating via typing and text is a sufficient way to identify intelligence (Assuming we go by the definition that intelligence is only logical question and answer). Use of visuals like video and photos of the people or listening to their voice should not really affect the intelligence of the person. Intelligence is not something that is depicted on a person’s face, however facial expressions might hamper or enhance signals of intelligence. I believe that expressions in a person’s eyes convey a state of mind and if the receiver is a good judge of signals can discount the obvious factors and read into the person’s expressions. If the receiver can tell that the person is being thoughtful or is distracted and/or does not really understand the conversation then video might be helpful. However, video and voice are ancillary, to identify intelligence textual communication will suffice. Assuming that certain types of communication is not meant for building social relationships, it might be that video and audio might actually hamper detection and reliability of signals of intelligence in those instances.
4.Paro…
Paro is an animal-robot and thus evokes reactions and feelings in people as they would when communicating with a real animal. Petting an animal is apparently therapeutic and relaxing. The article says that people with animals are in general happier or that strangers with animals are smiled at more. These theories are what makes Paro a success story when it comes to children and the ill. So the answer is yes, in this case deception is ok and in fact a good thing because it is improving the health and behavior of people. Using Paro for patients is a special case where deception is ok because the benefits of such deceptive signals outweighs the costs of deceiving. Given the condition of the people, I am not so sure it matters whether they know it is a robot or a real animal. Where children are concerned, many show as much love towards a stuffed dog as they do towards a real one, so using a robot dog should not be very different. In fact in children it might create more excitement. Adults depending on their mental state might not want to interact with something they know is not real but I think this number would be negligible.
Healthy children, on the other hand will know whether they are interacting with a real animal or a robot and depending on how old they are or how much they understand, their behavior might change. There would be no comparison to a real pet, because a real pet would have a mind of its own. A robot would probably be considered a real cool toy. For healthy and older children a real pet would probably be more therapeutic than a robot and thus preferred, but for children and elders with certain medical conditions, a robot pet is preferred so that it can be programmed to elicit certain behaviors in the patients and negative effects of allergies and biting are avoided.
The Imitation Game is far from being a good substitute for figuring out if machines can think because of the broad definition of the word ‘think’. Thinking, although intangible and does not require physical and voice representation, does require the ability to reason and understand, it is not merely giving answers to questions. Questions in an imitation game for computers are aimed at identifying if computers can think like men where reason and logic are required and it could be successfully used as a substitute in that regard. To imitate human conversation by following logic is not nearly as accurate as saying that the machine has intelligence or can think. The computer is given the logic by a human and so the reliability of using this game as a signal is lost there. If the signal of thinking and intelligence is restricted to questions and answers within the logic then yes, the game could be a reliable signal of intelligence, more like a reliable signal of whether the human is a good programmer and engineer! However, certainly we go by the definition that thinking and intelligence is way beyond just logical answers to consider the imitation game as a good signal of intelligence.
2.Weizenbaum…
The Rogerian or Person-centered therapy places the focus on the ‘client’ (not patient) to get him or her to talk about their ideas and feelings. Here, therapy and the role of the doctor is not to be prescriptive but let ting the patient talk about anything. Since Eliza is based on this theory, it does not require initiating a new chain of thought or generating ideas to give the client. ELIZA asks questions based on keywords in what the patient says and forming simple repetitive questions to hear the client’s point of view. Thus, people could communicate effectively with it knowing that it is substituting a therapist and so is not expected to have a normal social conversation with opinions, original thoughts, and retorts. This I think made people realize that ELIZA does not have intelligence of a human being because its responses are programmed.
I think that if a person does not know they are communicating with a chatbox, the conversation could be interesting and engaging but up to a certain point only, at which they may realize that the pattern of responses by the chatbot. Communicating with a chatbot whose model and structure is known, might have different results in a communication. The person communicating would know what responses/questions to expect and can direct the conversation the way he or she wants to get the desired outcome from the communication.
3.Donath…
Intelligence as defined by Webster is the ability to learn, understand, and reason all of which require mental ability. Because intelligence is intangible, communicating via typing and text is a sufficient way to identify intelligence (Assuming we go by the definition that intelligence is only logical question and answer). Use of visuals like video and photos of the people or listening to their voice should not really affect the intelligence of the person. Intelligence is not something that is depicted on a person’s face, however facial expressions might hamper or enhance signals of intelligence. I believe that expressions in a person’s eyes convey a state of mind and if the receiver is a good judge of signals can discount the obvious factors and read into the person’s expressions. If the receiver can tell that the person is being thoughtful or is distracted and/or does not really understand the conversation then video might be helpful. However, video and voice are ancillary, to identify intelligence textual communication will suffice. Assuming that certain types of communication is not meant for building social relationships, it might be that video and audio might actually hamper detection and reliability of signals of intelligence in those instances.
4.Paro…
Paro is an animal-robot and thus evokes reactions and feelings in people as they would when communicating with a real animal. Petting an animal is apparently therapeutic and relaxing. The article says that people with animals are in general happier or that strangers with animals are smiled at more. These theories are what makes Paro a success story when it comes to children and the ill. So the answer is yes, in this case deception is ok and in fact a good thing because it is improving the health and behavior of people. Using Paro for patients is a special case where deception is ok because the benefits of such deceptive signals outweighs the costs of deceiving. Given the condition of the people, I am not so sure it matters whether they know it is a robot or a real animal. Where children are concerned, many show as much love towards a stuffed dog as they do towards a real one, so using a robot dog should not be very different. In fact in children it might create more excitement. Adults depending on their mental state might not want to interact with something they know is not real but I think this number would be negligible.
Healthy children, on the other hand will know whether they are interacting with a real animal or a robot and depending on how old they are or how much they understand, their behavior might change. There would be no comparison to a real pet, because a real pet would have a mind of its own. A robot would probably be considered a real cool toy. For healthy and older children a real pet would probably be more therapeutic than a robot and thus preferred, but for children and elders with certain medical conditions, a robot pet is preferred so that it can be programmed to elicit certain behaviors in the patients and negative effects of allergies and biting are avoided.
posted by Pal at 6:36 AM
0 Comments:
Post a Comment
<< Home