Week1 Reading Notes

"Intelligent Machinery, A Heretical Theory":

• challenging the commonplace idea: machines cannot think ⇒ machines theoretically could simulate human thought (how?) ⇒ simulate human reasoning closely, making both mistakes and novel discoveries much like humans.

• Is simulated reasoning actual thought?

• Reflection: “Mary’s room qualia” + the chinese room experiment

• Learning by Experience: machines improve by learning from experience, evolving their behavior (like human education)

• Indexing Systems: a memory that tracks experiences with indexes ⇒ to recall past situations ⇒ future actions

• Crude Rules for Decision-Making becoming more sophisticated.

• pleasure principle: machine recognize favorable or unfavorable outcomes

• mimicking psychological motivations in humans.

• Randomness: diverse decision-making and reducing predictability ⇒ more adaptive learning

• machines could one day surpass human intelligence

• Humans will have to compete or collaborate with machines

• Already an reality these days?

• Ethical + Social Consequences: creating intelligent machines ⇒ opposition from intellectuals fearing job displacement + society fearing the rise of machines surpassing human control

Plant - Zeros + One - 88-95

• Irony in Human superiority: Man is superior ⇒ Turing: machines would challenge man’s superiority

• Creating machine as slave ⇒ performing predictable tasks given by creators

• Backfire? 

• Process become routine ⇒ machine can be programmed to handle them autonomously ⇒ reducing human control

• Human masters would protect their roles ⇒ keeping their work in mystery + complexity

• Chatbot interactions: imitation + authenticity ⇒ what it means to be human in the face of increasingly intelligent machines?

• Trials:

• Eliza: psychotherapist using clever tricks in the imitation ⇒ no contribution + contradiction to human messages

• Parry: neurotic male

• Julia: “agent more interesting than Eliza”

• Adept at detecting and deflecting sexual advances

• Turing: computing machines doing “errors” other than what they were instructred ⇒ mistakes are a most intelligent move to be like human ⇒ differences b/w failures and redusals to be bound by them?

• Perfection != success

Bratton - AgueroyArcas - The Model Is the Message

• Bratton: does not believe LaMDA is conscious as Lemoine believes

• Sentience : need to define in both intelligent and conscious aspects

• Apply existing philosophy vs. invent new concepts and terms to make sense of technologies before

• Not interior subjectivity to human, but projecting personhood onto a pre-scripted chatbot (the ELIZA effect)

• Graziano: consciousness = mind emphathetically modeling other minds

• Subjectivity: objectifying one’s own mind

• Mary’s room: the scientific knowledge vs. the actual experience

• Conversations: “synthetic personalities”

Oct 8 2024 Discussion in class:

• The smartest human would design the smartest AI in the world?

• Sentience: the emotional/conscious  intelligence(less tangible feelings), while intelligence is capacity to learn(more rational)

• Sefl- swareness of “who you are” - does the AI know it’s an AI?

• Mirror stage of young children - able to recognize oneself in a mirror in the growth of kids

my short definition:

Intelligence would go beyond simply storing and retrieving information. It would involve understanding and responding to both expressed and hidden emotions. True intelligence is rooted in empathy, the capacity to connect with others’ feelings and adjust accordingly. This would include intuition, perception, imagination, and a deep understanding of human emotions, and mimicing warmth or human-like reactions.

keywords:

embracing, connectivity, sensitivity

illustrations:

embracing