 | This is the sandbox page where you will draft your initial Wikipedia contribution.
If you're starting a new article, you can develop it here until it's ready to go live. If you're working on improvements to an existing article, copy only one section at a time of the article to this sandbox to work on, and be sure to use an edit summary linking to the article you copied from. Do not copy over the entire article. You can find additional instructions here. Remember to save your work regularly using the "Publish page" button. (It just means 'save'; it will still be in the sandbox.) You can add bold formatting to your additions to differentiate them from existing content. |
Can Chat GPT think?
editAs Chat GPT produces work, many question if it is actually thinking. Many philosophical tests can be applied to Chat GPT to understand if it can think. One of these tests is the Turing Test made by Alan Turing, which was developed to see whether a machine can think.[1] The concept of this test is to see whether a machine can fool an investigator into thinking that he is a human and if it can, then the machine can think.[1] According to the Turing Test, if Chat GPT can think it would be able to have a conversation because that requires one to have to reflect on what to respond, which would prove that Chat GPT is thinking.[1]
The Turing Test has many objections. The first objection that arises is if the test is human-centric.[2] The Turing Test could be forcing Chat GPT (or another machine) to think like a human, so even if Chat GPT fails the test it still might be able to think, just not the same way as a human.[2] However, Turing responded to this objection and states that he was only trying to offer a sufficient condition (a condition or set of conditions that will ensure the occurrence of a specific event or outcome) and not a necessary condition (a condition or set of conditions that must be met in order for a specific event or result to happen) for a machine to think.[2] So, if Chat GPT does not pass the test, we shouldn’t conclude that it couldn’t think because it is not a necessary condition for it to think.[2]
The second objection is from Lady Lovelace which states that for Chat GPT to think it needs to use originality or creativity and if Chat GPT can pass the Turing Test it would only mean that it has good programming.[3] There are two ways of interpreting this argument. The first, Chat GPT (or another machine) would need to do something original to prove that it can think.[3] Turing’s argument to this objection is that even humans do not always do something new, but we are considered as being able to think.[3] The second way of interpreting this argument is that Chat GPT only does things that its programmers have told it too, so the Turing Test wouldn't be testing the machine, it would be testing the programmers.[3] Turing’s counter-argument is that humans are also programmed to do certain things, we just consider it learning.[3] So Chat GPT would still be considered able to think even after Lady Lovelace’s objection.[3]
The final objection that arises is the argument from consciousness.[4] Many believe that for something to think they have to have consciousness, however, the Turing Test only tests how Chat GPT would behave, not whether it is conscious or not.[4] So, according to this objection, the Turing Test would not accurately test if Chat GPT can think or not because it does not test its consciousness.[4] One argument that is brought up from this is that what is happening inside of Chat GPT is unknown.[4] However, Turing explains that if we believe this argument, we might fall into believing solipsism.[4] He states that what is going on in Chat GPT is unknown, just like what is going on inside another human is unknown.[4] Another argument that arises from the argument of consciousness is Susan Schneider’s argument. She states that when diagnosing a medical illness, doctors often use more than one test, so while trying to figure out if Chat GPT (or another machine) is conscious or not, one should do the same.[4] She makes suggestions on what tests to use, including the ACT (the AI Consciousness Test).[4] The ACT tests if a machine has developed its own views and experiences.[4] In order to use the ACT, Chat GPT (or another machine) must have no previous programming or 'knowledge' about consciousness.[4] Then, the test consists of questioning Chat GPT about how it feels to exist, how it experiences certain things, etc..[4] Schneider also suggests the ACT as a sufficient condition and not a necessary one to consciousness.[4] In other words, if Chat GPT doesn’t pass it does not mean that it does not have consciousness.[4]
- ^ a b c Kind, Amy (2020). Philosophy of Mind: The Basics. Abingdon, Oxon and New York, NY: Taylor & Francis Group. pp. 98–100. ISBN 978-1-315-75090-3.
- ^ a b c d Kind, Amy (2020). Philosophy of Mind: The Basics. Abingdon, Oxon and New York, NY: Taylor & Francis Group. pp. 100–102. ISBN 978-1-315-75090-3.
- ^ a b c d e f Kind, Amy (2020). Philosophy of Mind: The Basics. Abingdon, Oxon and New York, NY: Taylor & Francis Group. pp. 102–104. ISBN 978-1-315-75090-3.
- ^ a b c d e f g h i j k l m Kind, Amy (2020). Philosophy of Mind: The Basics. Abingdon, Oxon and New York, NY: Taylor & Francis Group. pp. 104–106. ISBN 978-1-315-75090-3.