Week 2 delves into early attempts to replicate human conversation through chatbots, such as ELIZA. The students will explore the rationale behind these developments and the ethical dimensions of mimicking human interaction. By investigating historical examples like Weizenbaum's ELIZA, students will gain insights into the technological and social ambitions that drove early AI researchers.
There is, of course, no generally accepted theory of "intelligence"; the analysis is our own and may be controversial. - Marvin Minsky, 1960
- Human-Machine Interaction: The interaction and communication between humans and computer systems.
- ELIZA: An early AI program that simulated a Rogerian psychotherapist, created by Joseph Weizenbaum.
To Read (Before Class)
- Tung-Hui Hu’ "A Prehistory of the Cloud: Virtualization and the Architectures of Networked Computing" (2019) - Read the introduction and chapter 1 A Prehistory of the Cloud
- Pay attention to the “The Graft” - we are going to graft AI onto existing technologies, it will not be developed in a vacuum.
Natural Language Communication Between Man and Machine (1966)
- "Artificial Intelligence: A Guide to Intelligent Systems" (Chapter 1)
In Class Assignment
Break out into small groups (3 to 4 students) and spend some time talking with ELIZA. Answer a few of the questions below & record your answers in this week’s journal.
- What is the impact of users believing ELIZA's responses were coming from a human being?
- Compare the simplicity of ELIZA's design compared to modern AI systems
- What are some of the early ethical boundaries the researchers stumbled over?
- Do you think this open up concerns over deception, trust, and reliance on machine-driven communication?