IAI Chapter A Philosophical Aspects
Questions about the lecture 'Introduction to Artificial Intelligence' of the RWTH Aachen Chapter A Philosophical Aspects
Questions about the lecture 'Introduction to Artificial Intelligence' of the RWTH Aachen Chapter A Philosophical Aspects
-
- 1 / 28
-
Cartes-fiches
What is a problem asking about AI?
[philosopical]
Questions like “can machines think?” do not define thinking
What did Steven Hawking said?
[development.philosopical]
“Full AI could spell end of human race… Due to re-design speed compared to slow biological evolution”
What did Elon Musk said?
[development.philosopical]
“Biggest existential threat is probably AI. We would be summoning demons”
What is Moore’s law?
[development.philosopical]
Computers become more and more powerful
What are the aspects by Kurzweil?
[technological singularity.development.philosopical, 2]
1. AI will necessarily progress exponentially
2. In the “end” exists super intelligence (SI) making humans obsolete
What are the difficulties?
[technological singularity.development.philosopical, 4]
1. Intelligence is more than quick thinking
2. Software development lags behind hardware development
3. Unsolvable problems (even for super humans) exist
4. Why should human intelligence be turning point evolving to SI
Where is big progress made?
[development.philosopical, 4]
1. Image and speech recognition
2. Question-/Answering system // Watson
3. Game playing // Chess, poker, go
4. Autonomous driving
Where is some progress made?
[development.philosopical, 2]
1. Making rational decisions
2. Reaction appropriately in unforeseen situations
Where is little progress made?
[development.philosopical, 4]
1. Ethical decisions
2. Emotions
3. Intuition
4. Consciousness
What should we be worried about and what not?
[development.philosopical]
Not worried about Terminator but maybe about drones
What is the definition?
[weak AI.philosopical]
Machines act if they were intelligent // Mostly in this course
Why is it impossible?
[weak AI.philosopical, 4]
1. Only do what they are told
2. Can not make ethical decision
3. Can not have emotions, humor, etc
4. Mathematical argument via Gödels incompleteness result
What are the characteristics?
[gödel incompleteness result.weak AI.philosopical, 3]
1. For all formal system F with arithmetic sentence G(F) exists
2. G(F) consistent and true in intended interpretation
3. F !|- G(F)
What are the inferences?
[gödel incompleteness result.weak AI.philosopical, 2]
1. Computers are formal systems, do not recognizes truth
2. Humans are no formal systems, do recognizes truth
What is the definition?
[strong AI.philosopical]
Machines have real (human like) conscious mind
What is the reaction from Turing?
[strong AI.philosopical]
“Even the question makes no sense. Who know if other humans are conscious? …although It seems pretty likely”
What is the setting?
[searle’s chinese room.strong AI.philosopical, 3]
1. Thought experiment
2. Room with a person (english), rule book (english), empty pages and pages with strange symbols
3. Strange symbols correspond to Chinese
What is the procedure?
[searle’s chinese room.strong AI.philosopical, 4]
1. Person receives piece of paper with strange symbols trough slot
2. Strange symbols are in rule book connected to instructions
3. Person writes additional strange symbols
4. Piece of paper is passed back
What is the resulting questions?
[searle’s chinese room.strong AI.philosopical]
“Does the room understand Chinese?”
Which is the set of objects not understanding Chinese?
[searle’s chinese room.strong AI.philosopical, 3]
1. Person
2. Paper
3. Rule book
What is the conclusion?
[searle’s chinese room.strong AI.philosopical, 2]
1. If each object in set has no conscious mind then composition of objects can not be conscious ever
2. Hence “Chinese room” does not understand Chinese
What is the definition by Koch?
[functionalism.searle’s point of view.strong AI.philosopical, 3]
1. System and its parts are determined by I/O behavior
2. Brain work depends on topology and I/O behavior of neurons
3. In principle human brain could at least be simulated
What can be deduced from Searle as anti-functionalist?
[anit-functionalism.searle’s point of view.strong AI.philosopical, 3]
1. Consciousness is an intrinsic property of brain
2. Then it is an epiphenomenon (i.e. has no causal effect)
3. But how can conscious perceptions as pain be explained
What is the setting?
[levesque’s summation room.strong AI.philosopical, 4]
1. Similar to Searle’s Chinese room
2. Input consists of twenty 10-digit number
3. Output is 12-digit number corresponding to sum of input numbers
4. Rule book
What are the possibilities?
[rule book.levesque’s summation room.strong AI.philosopical, 2]
1. Lookup table
2. Algorithm
What are the characteristics?
[lookup table.rule book.levesque’s summation room.strong AI.philosopical, 2]
1. Searle is right, person can not add
2. Impossible due to size
What are the characteristics?
[algorithm.rule book.levesque’s summation room.strong AI.philosopical]
We would probably say person can add
What is an example?
[gödel incompleteness result.weak AI.philosopical]
“Susan cannot consistently assert that this sentence is true”
