Assignment AI_ week 1_ Komalasari _2681483783

Nama : Komalasari

Nim : 2681483783

Question :

• How would you explain the difference between narrow AI and general AI to someone with no technical background?

• What was the significance of the Dartmouth Workshop in 1956?

• In your opinion, is the Turing Test still a valid measure of machine intelligence? Justify your answer.

What is the “AI effect,” and can you provide a modern example?
Answer :

1.The Difference Between Narrow AI and General AI

Narrow AI is AI designed to do one specific task only. It is extremely skilled in its domain, but cannot do anything outside of it.

Simple analogy: imagine a calculator that is brilliant at doing math, but cannot cook, drive, or write poetry. That is narrow AI.

Examples of narrow AI:

• Voice assistants (Siri, Google Assistant) — only answer questions and execute commands
• Netflix recommendations — only suggest movies
• Email spam filters — only sort out junk mail
• Self-driving cars — only drive

General AI is AI that has human-like cognitive abilities across the board. It can learn, reason, and perform many different tasks, just like a human.

Simple analogy: imagine a versatile human who can cook, drive, write poetry, learn new things, and switch from one task to another without needing to be reprogrammed.

Key differences:

Key point: All AI that exists today is narrow AI, even the most advanced systems like ChatGPT. General AI remains a theoretical concept and has not yet been achieved.

 2.The Significance of the Dartmouth Workshop in 1956

The Dartmouth Workshop, held over approximately eight weeks in the summer of 1956 at Dartmouth College in New Hampshire, has several key significances:

• • Birth of the Term “Artificial Intelligence”
    John McCarthy and three other scientists—Marvin Minsky, Nathaniel Rochester, and Claude Shannon—wrote a proposal for this conference and coined the term “artificial intelligence”. The term was deliberately chosen to “nail the flag to the mast” and focus attention on the goal of making machines behave intelligently. McCarthy later wrote the proposal in August 1955 that became the source of the term.
  • Establishment of AI as a Field of Study
    The conference symbolically established AI as a discipline in its own right. McCarthy himself stated that “the symbolic role of the Dartmouth Summer Workshop on Artificial Intelligence in establishing AI as a field of research was more important than the specific results obtained at the meeting”. Previously, research on intelligent machines was scattered across fields like mathematics, psychology, and engineering; Dartmouth coalesced them into a single coherent field.
  • Paradigm Shift from “Automata Studies” to “AI”
    Claude Shannon preferred the term “automata studies,” while Allen Newell and Herbert Simon used “complex information processing.” However, the term “artificial intelligence” proposed by McCarthy ultimately won out and became the official name of the field. This naming choice also carried consequences: from that point on, AI has always been compared to human intelligence—a comparison that has been “both a blessing and a curse” for the field’s development.
  • Overconfidence and a Complicated Legacy
    The participants were incredibly optimistic that they could “solve the problem of machine intelligence in a single summer”. McCarthy even hoped that “by summer 1956 I will have a model of such a machine fairly close to the stage of programming in a computer”. This optimism did not materialize, but it became a recurring pattern in AI history: cycles of hype and disappointment.

An Enduring Legacy
Although the conference failed to achieve its ambitious goal of building a thinking machine in one summer, the Dartmouth Workshop “turned AI research into a scientific discipline of its own”. Dan Rockmore, professor of math and computer science at Dartmouth, stated: “We can honestly say that AI was born here at Dartmouth

3.In my opinion, the Turing Test remains historically and conceptually relevant, but is no longer sufficient as a sole measure of machine intelligence. Here is my justification:

Why the Turing Test Remains Valid:

1. Focus on Behavior, Not Mechanism
   Turing cleverly shifted the question from “can machines think?” to “can machines behave as if they think?”. This is a functionalist approach that remains useful: we judge AI by what it does, not how it does it.
2. The Indistinguishability Test
   The core of the Turing Test—a machine’s ability to deceive a human into not distinguishing it—remains a real challenge. Even modern AI like ChatGPT can still be identified as a machine in extended conversations, though it is becoming increasingly difficult.
3. Practical Relevance
   In applications like customer service chatbots or virtual assistants, the ability to “pass” as human is indeed a practical goal.

Why the Turing Test Is No Longer Sufficient:

1. Intelligence ≠ Ability to Deceive
   The Turing Test measures the ability to imitate humans, not intelligence itself. A bird that mimics human speech does not become human. An AI that memorizes human conversation patterns can pass this test without true understanding.
2. The “Large AI” Era Changes the Landscape
   Large language models (LLMs) are trained on billions of human texts. They can mimic human conversational style extremely well—but this demonstrates pattern imitation more than understanding. Passing the Turing Test is now easier because AI is trained on human data, not because AI is truly “intelligent”.
3. It Ignores Other Aspects of Intelligence
   The Turing Test only tests textual communication. It ignores:

• Spatial intelligence (navigation, object manipulation)
• Emotional intelligence (genuine empathy)
• Original creativity (not just recombination)
• Continuous learning in the real world
• Consciousness and self-awareness

1. The “Chinese Room” Problem
   Philosopher John Searle argued that a system manipulating symbols (like a computer) could pass the Turing Test without genuinely “understanding” anything—like a person in a room following rules to answer in Mandarin without understanding the language.
2. Turing Himself Doubted It
   Turing predicted in 1950 that within 50 years, computers would fool 30% of humans for 5 minutes. This prediction has nearly been met, but he himself called the Turing Test an “imitation game”—not a definition of intelligence.

Conclusion:

The Turing Test remains valuable as a minimal criterion and a historical milestone, but is no longer adequate as a comprehensive measure of machine intelligence. We need a richer evaluation framework: generalization ability, causal understanding, continuous learning, and perhaps consciousness. 

Status : terselesaikan 100%