Unit 4: Case studies

Activity 1: Review

Work in pairs. Without referring to the course website or any notes, jointly create a list the technical terms related to expert systems that you covered so far.

Activity 2: Evaluation of expert systems for adjective order

The images below are representations of expert systems created by course participants. Each system has both pros and cons. This activity focuses on identifying areas for improvement, though.

Evaluate each of the expert system designs by considering the following:

1. Does the system provide expert answers for each instance?
2. Can the system be converted into a program without further analysis?

Activity 3: Evaluation of expert systems for two future forms

The images below are representations of expert systems created by course participants. Each system has both pros and cons. This activity focuses on identifying areas for improvement, though.

Evaluate each of the expert system designs by considering the following:

1. Does the system provide expert answers for each instance?
2. Can the system be converted into a program without further analysis?

Activity 4: Evaluation of expert systems for stressed syllables

The images below are representations of expert systems created by course participants. Each system has both pros and cons. This activity focuses on identifying areas for improvement, though.

Evaluate each of the expert system designs by considering the following:

1. Does the system provide expert answers for each instance?
2. Can the system be converted into a program without further analysis?

Activity 5: Cheating with ChatGPT

Read.

Following the release of ChatGPT online, some students have submitted AI-generated work without declaring the use of AI. Professors on reading the work can soon realize that the submission was not written by a student. However, the professor needs to use time to read and assess whether the work was student-created or AI-generated.

Texts generated by ChatGPT share some characteristics. These include:

1. stating that it was trainied on data up to September 2021
2. reiterating the input
3. using bulletpoint lists
4. starting bullet point items with a noun phrase written in bold followed by a colon
5. using accurate spelling
6. selecting American rather than British spelling
7. using accurate grammar
8. not making predictions about the future
9. plagiarizing texts
10. hallucinating, i.e. making up facts
11. longer than necessary explanations (because it is unaware of the knowledge of the reader)

Work in pairs. Discuss how to create an expert system to automatically detect whether a text was generated by ChatGPT.

Activity 6: Visualing expert system: Practice

Work in pairs or threes. Create a flowchart for the ChatGPT-generated text detection expert systems. Submit your work on ELMS.