AI Literacy Pro? Take the 5-question challenge. Score: 0–2 = getting started; 3–4 = AI-aware; 5 = AI literacy pro.

Q1. A system learns from labeled photos so it can classify a new photo as showing a bagel or a donut. Which choice describes the machine-learning process, rather than the task?
- Classifying the new photo
- Learning patterns from examples to classify a new photo
- Taking a photograph
- Choosing the labels after classification
Answer: Learning patterns from examples to classify a new photo[[cite:1]][[cite:2]]. Explanation: Classification is the task; learning from examples is the process the system uses to do it.
Q2. A text-generating AI gives a confident answer with a citation. What is the best next step before relying on or sharing it?
- Assume it is true because it sounds confident
- Check the claim against reliable sources
- Trust it if the answer is detailed
- Treat the citation as proof without checking
Answer: Check the claim against reliable sources[[cite:3]][[cite:4]]. Explanation: Generative AI can produce confident but inaccurate content, including citations, so its claims need checking.
Q3. A team notices that an AI system's outputs reflect stereotypes. Which explanation best matches how bias can enter a system?
- Only through the wording of a user's prompt
- Through data, selection and labeling, model training, or developer instructions
- Only after a system has been deployed
- Bias cannot affect AI outputs
Answer: Through data, selection and labeling, model training, or developer instructions[[cite:5]][[cite:6]]. Explanation: Bias can enter at several stages, and efforts to reduce it do not guarantee its absence.
Q4. You are considering entering sensitive personal details into an AI prompt. What is the most responsible first question to ask?
- Whether the prompt is long enough
- What data is necessary and how it may be used
- Whether the AI can make the details sound clearer
- Whether the details will improve the response's style
Answer: What data is necessary and how it may be used[[cite:7]][[cite:8]][[cite:9]]. Explanation: Privacy-aware use includes considering what information is needed and how collected information may be used.
Q5. An AI tool recommends an option for a consequential decision. Which approach best reflects responsible use?
- Accept the recommendation automatically because the system is consistent
- Check the output, consider possible harms and fairness, and use human judgment
- Ignore privacy because the decision is important
- Assume a clear prompt guarantees the recommendation is correct
Answer: Check the output, consider possible harms and fairness, and use human judgment[[cite:10]][[cite:11]][[cite:12]]. Explanation: Human oversight matters, especially for consequential decisions; a system's output should not be accepted automatically.

AI Literacy Pro? Take the 5-question challenge. Score: 0–2 = getting started;

Explorer des sujets liés