Artificial intelligence (AI) builds systems that perform tasks associated with human intelligence. In practical data work, it can help with model building, statistical analysis, information retrieval, and Python code generation. section 10.16
Generative AI is especially useful as a fast assistant, provided its output receives human review.
- Use tools such as Claude, ChatGPT, Gemini, or Grok to suggest Python snippets, functions, libraries, and reusable coding patterns. section 10.16
- Treat generated answers as provisional: AI can produce confident mistakes, misidentify people, or return flawed analysis. section 10.16
- Check how an answer was produced instead of outsourcing an entire CSV analysis without reviewing its reasoning. section 10.16
- Label questionable claims as hypotheses rather than facts, and remain transparent about uncertainty. section 10.27
A speedy assistant that gets names wrong
A Gemini assistant was asked to identify a person at a video timestamp and repeatedly supplied incorrect names. An AI reverse-image search also failed, while manual investigation identified the person. Verification prevented a false attribution. section 10.16






















