Lecture Recall is a generative-AI study assistant—a custom GPT inside UPenn’s ChatGPT Edu workspace—that lets students revisit, search and quiz themselves on recorded class sessions, slide decks and practice exams. This tool is currently available in selected pilot courses only.
Before You Start
You will need the following in order to use Lecture Recall:
- Active pennkey (
username@upenn.edu) - Enrollment in a course that is running the pilot.
How to Use the Service
- Log in with SSO: Go to chatgpt.com and log in using your pennkey@upenn.edu address. You can tell you are in the UPenn ChatGPT Edu by the Penn shield in the bottom left corner of the interface. If it's not there:
- Log out, and Log back in.
- Choose “Sign in with SSO” and enter
username@upenn.edu
- Navigate to your Course Lecture Recall Link: The link will be in your Canvas Site, distributed by your teaching team. If you cannot find it, ask your TA or professor.
- Pin your course GPT: click the star to keep the course-specific GPT link in the sidebar for one-click access.
- Ask targeted questions: e.g.: "In the lectures, what did prof say about..." See the Queries Section below for more details.
- Submit feedback for the teaching team: Tell the bot "Submit Feedback {your feedback | e.g.: this answer showed the wrong formula}."
- Cross-check facts: treat the bot as a study aid, not an official source—always validate critical information by following the video links, or follow up with your TA or professor.
Anonymized Usage Data
Our intent is to understand how the tool is used, so that we can continually improve the user experience, while preserving your privacy. Currently, Lecture Recall collects queries, but it anonymizes the username, so we can tell a query was sent, but not who sent it. We store anonymized names like you see in Google Apps, such as "balcony pear" and "debris blanket."
If you think this is too much or too little, please let us know directly in the bot (e.g.: "Submit Feedback - Anonymizing Usage data is a great idea, but here's what I think...")
Queries
Let's walk through some of the particulars of working with the bot. Keep in mind that the teaching team does often customize the instructions and files available to the bot, so behavior might vary. However, there are some fundamentals that should help you get the most out of your review time.
- Oauth (Authentication): The first time that the bot sends your query to our database, you will be asked to confirm the transmission. Select Yes Always for the most convenient experience.
- Know what data is being used: Although we instruct the bot (quite forcefully!) to always reference the course content, it does not always do so.
- Talking to apps.wharton... -> Sending your query to our database containing lecture transcripts. This will return the closest semantic matches with bookmarked links to the video
- Searching my knowledge... -> Searching any files that the teaching team has uploaded. Sometimes the search can occur in milliseconds, so you might not have time to notice the wording of the action.
- Ask explicitly for what you want:
- If you want to search the lectures -> "What do the lectures say about {my question}?"
- If you prefer the course notes -> "What do the course notes mention about {my question}"
- Some classes also arrange their lectures and course notes by a simple index. If they do, then you can reference the course session # as well -> "In the first class session, what did prof say about {my question}?"
Frequently Asked Questions
Q: Can Lecture Recall see the graphs or images I upload?
A: No. Describe the graph or paste the underlying data; the bot will interpret it textually.
Q: The bot gave an answer that conflicts with the lecture. What do I do?
A: Flag it immediately with “Send feedback: [description of issue]” or open a support ticket with screenshots so we can retrain the model.
Q: I’m enrolled but the GPT link says “Access Denied”.
A: Confirm you’re logged in with your username@upenn.edu and not a personal Gmail. If the issue persists, clear your browser cache or try incognito mode, then contact support.
Questions?
Contact: Submit a ticket or email support@wharton.upenn.edu
Support Hours: Mon–Fri, 9 a.m.–5 p.m. ET
Response Time: within 1 business day