引導式測驗 · 中等 等級
ELIZA 效應:為什麼我們將聊天機器人擬人化測驗
Learn the ELIZA effect’s history, conversational cues that invite anthropomorphism, and why impressions differ from evidence.
問題 1 的 8
A program reflects a user’s words as a question, and the user says it “really understands me.” What idea does this illustrate?
預覽每個問題8 問題
當您參加測驗時,就會出現答案和解釋。
- A program reflects a user’s words as a question, and the user says it “really understands me.” What idea does this illustrate?
- The ELIZA effect: attributing more human understanding to a program than its behavior establishes
- A verified measure of the program’s emotional experience
- A proof that pattern matching cannot produce conversation
- A benchmark showing the program has human-level reasoning
- Which historical feature characterized Weizenbaum’s original ELIZA program?
- A neural network trained on internet-scale dialogue
- Pattern matching and scripted text transformations
- A web-search index of current news
- A camera system that reads facial expression
- A chatbot uses a person’s name and recalls a saved preference. What can the user reasonably infer?
- The system has feelings for the user
- The stored preference proves human-like memory
- The system may be using stored context, but the cue alone does not establish a personal relationship
- The response was written by a human operator
- Why is friendly, fluent wording not enough to show that a chatbot understands a user’s full situation?
- Every fluent answer comes from a human
- Only neutral wording can communicate facts
- The ELIZA effect proves all chatbot responses are false
- Conversational style is a social cue, not independent evidence of complete understanding
- How did different mental-state attributions relate to advice-taking in the 2025 experiment described?
- Intelligence and experience attributions had different relationships, and the consciousness result depended on the analysis
- Every mental-state attribution had the same positive relationship
- The study established that chatbots have conscious experience
- The task measured all real-world consequences of anthropomorphism
- A conversation feels private because a bot responds warmly. What should the user check before sharing sensitive details?
- Whether the bot uses emojis
- The product’s privacy practices and what information it stores or uses
- Whether the response is written in first person
- How quickly the bot replies
- Which design choice can help users understand a chatbot’s role?
- Give it a human photo without disclosure
- Describe every answer as a personal opinion
- Identify it as software, explain memory controls and provide a human route
- Hide whether conversations are stored
- What should an evaluation of anthropomorphic interfaces measure separately?
- Only whether users say the bot is friendly
- Only the model’s number of parameters
- Whether the interface uses a human name
- Perceived empathy, trust, disclosure and reliance as distinct outcomes