top of page
Writer's pictureKeith Hannaleck

Getting ChatGPT to Do My Job For a Day

With so much talk about ChatGPT replacing jobs (which is in fact a true concern). I wondered if ChatGPT could do my job for a day. (Mostly because it was the weekend and it’s more challenging to work on the weekend.) I asked it to do a couple of tasks. Here are some results.

Task 1: Create a test that tests job skills

I asked ChatGPT to create some questions to test a job candidate’s job skills. It generated an answer in a matter of seconds:

Sure! Here are some sample questions you could use to test a job candidate’s proficiency at customer service:

  1. How do you handle a customer who is upset or dissatisfied with the product or service they received?

  2. Can you provide an example of a time when you went above and beyond to help a customer?

  3. How would you deal with a customer who is speaking loudly or rudely to you?

  4. Describe a time when you had to handle multiple customer requests simultaneously.

  5. How do you stay calm and composed when dealing with difficult or angry customers?

  6. What do you think are the most important qualities for someone in a customer service role?

  7. How do you handle a situation where you are not able to provide a customer with the solution they were hoping for?

  8. Have you ever dealt with a language barrier when speaking with a customer? How did you handle it?

  9. How do you keep yourself organized when handling multiple customer requests or issues?

  10. Can you provide an example of a time when you successfully turned a dissatisfied customer into a loyal one?

Task 2: Create an answer key that can be reviewed using AI

Perhaps I didn’t ask the right question. When I asked ChatGPT to create an answer key to a job skills test using AI, it came up with this response:

To have AI check the answers to these questions in a standardized way, you can follow these steps:

  1. Develop a set of criteria for evaluating each question. This could include factors such as the clarity and completeness of the answer, the use of relevant examples, and the candidate’s ability to communicate effectively.

  2. Train a natural language processing (NLP) model on a large dataset of customer service responses to the questions you have developed. This will enable the model to recognize and understand common patterns and language used in customer service responses.

  3. Use the trained NLP model to automatically evaluate the candidate’s responses to the questions. The model will analyze the text of each response and score it according to the pre-defined evaluation criteria.

  4. Set a threshold for acceptable scores for each question. Responses that score below this threshold can be flagged for review by a human evaluator.

  5. Review the results of the automated evaluation and make any necessary adjustments to the evaluation criteria or the threshold scores.

Conclusion:

When I asked ChatGPT to create a job skills assessment as well as a method to evaluate the assessment using AI, it replied with well-written responses. However, none of it was material I could use because the material was far too general. The issue was also my inability to word my task to get the answer I really wanted: to get ChatGPT to do my homework. What I learned was that it may take some skill to learn how to get ChatGPT to read my mind and do my creative thinking so I can take an afternoon off.

If you enjoyed this article, subscribe to make sure you don’t miss the next post!

0 views0 comments

Comments


bottom of page