Community Discussion · Policy

My Test Results: This AI Literacy Tool Reveals Not Dumb AI, But Our Poor Prompting Skills

Feng sirAug 92026/08/09 331 views

Last week, after finishing a model inference run in the lab, I casually clicked on an AI literacy assessment link submitted by a student. The tool itself is a chatbot that has you interact with it for 20 to 40 minutes, then generates an evaluation report based on your conversation habits. I initially thought it would just be a formality, but after running through it, I sat in front of my screen stunned for quite a while.

Policy

2 replies

Ctrl + Enter to reply

A DeerAug 9

Lol, isn't this exactly what happened when I interviewed a candidate last week... Their resume said 'Proficient in Python,' but as soon as I asked about project details, they started being vague. It's the same principle when talking to AI: if you don't provide context, it can only give you generic responses. I've had this feeling recently using WorkBuddy to screen resumes too—everyone is stuffing keywords, but there are very few people who can actually clearly explain what they've done.

Is Operator Fusion Done?Aug 9

Could you DM me that link to the evaluation tool? I've recently been trying out Qwen-7B as a beginner, and I feel that the way questions are asked really makes a big difference. I've encountered similar situations before—asking with just keywords resulted in all generic answers, but later switching to complete questions with context made the answer quality completely different.