Final Exam Preparation: Human Behavior Week 14

End of Semester Housekeeping

Sign-in for attendance.
Complete student feedback for the subject (unit) and the teaching (instructor).
- Email reminders were sent on May 29th.
- Feedback is used to refine the unit and improve subjects.
- Teaching feedback impacts instructor's job prospects.

Format: Pen and paper.
Date and Time: June 19th at 2 PM.
Location: Find the room via the my exam timetable.
Content:
- 40 multiple-choice questions.
- 30 simple and scenario-based questions focused on lecture content.
- 10 questions outlining an AI academic summary evaluation task.
Duration: 1 hour.
Allowed Materials: Handwritten or printed notes (multiple sheets, less than a small book).

Prioritize: Core lecture content.
Steps:
1. Watch all lecture videos.
2. Read associated weekly readings.
3. Review tutorial slides and email records (weekly activity submissions) if time permits.
Caution: Excessive notes during the exam may waste time.
Recommendation: Writing summary notes is excellent for memory scaffolding and understanding.
Access to Content: Unit content is typically available until the end of the exam period (including catch-up exams).
Personal Recommendation: Download all lecture materials and tutorials for future reference.
Deferred Exams: Extensions for access to the site can be requested.

H5P Game: Practice with five example questions.
Full Practice Exam:
- 30 questions.
- Timed.
- Covers full subject content.
- Located on the tutorial repository page.
- Only one attempt is allowed.
- Correct answers supplement research participation credit points.
Behavior 2018 Final Exam Tutor (Microsoft AI Copilot):
- Lower stakes option.
- Coaches on why correct answers are correct.
- Access via student emails (nine dots > Microsoft 365 Copilot).
- Custom Copilot agent chatbot using lecture transcripts.

Activate Microsoft 365 Copilot via student emails.
Install the AI created for the unit via the link in week four tutorial repository.
Example prompt: "Please give me one multiple-choice question based on week two content."

AI works by pattern recognition, using Bayesian statistics to determine the most likely next word.
Large language models tokenize words and analyze their weightings and connections.
AI doesn't "know" what it's saying; it predicts the likelihood of token sequences.
Fact-check AI-generated content, as it can produce incorrect answers.
Analogy: AI is like completing a jigsaw puzzle using pattern recognition, not intelligence.

Task: Evaluate a 500-word AI-generated summary of a psychology journal article.
Answer 10 multiple-choice questions about the summary's accuracy by cross-referencing it with the original article.
The original journal article will be provided in full during the exam.
Recommendation: Read or watch the article beforehand.

Custom Copilot AI chatbot designed to generate brief, structured academic summaries with intentional errors.
It asks multiple-choice questions about the summary, modeled after exam questions.
Caution: Automated responses may be janky; challenge the AI if you believe it's wrong.
Access via the second link in the week four tutorial repository.
The journal article for the AI academic summary evaluation task is "Mined Over Milkshakes."
Recommendation: Use Chat GPT or other AI with the original article to generate a 500-word summary with internal errors for personal practice.

Due to limited research studies, alternative tasks are available to earn missing credit points.
Options:
1. Practice Exam: Every correct answer gives 0.5 credit points (maximum 15 points from 30 questions).
2. Lecture Reflections: Approved lecture reflections contribute 5 credit points each (maximum 6 reflections for 30 points).
Lecture Reflection Guidelines:
- Focus on a new lecture week each week.
- Duplicate reflections will not be credited.
- Do not use AI.