Set Up ChatGPT for Your Studies

Optimize Your Work Environment

Having a ChatGPT account is a good start; configuring it for your academic reality is what transforms your weeks. A student who has filled out their profile, organized their conversations, and selected the right models can achieve in three exchanges what another student spends twenty minutes searching for. The full setup takes about an hour, done just once, and pays for itself during the first week of classes.

Choose the Right Subscription

The plan you use determines which models you can access, which in turn defines the type of work you can entrust to the tool.

Before you pull out your credit card, check if your university offers ChatGPT Edu. Many institutions include it in tuition fees, providing full access to advanced models and enhanced privacy protections.

Custom Instructions: Your Student Profile

Repeating “I am a second-year law student, respond in academic French” at the start of every new conversation is a pure waste of time. Custom Instructions, found under Settings → Personalization → Custom Instructions, solve this problem once and for all: ChatGPT applies this context to all your interactions.

The first field describes who you are. Fill it with your actual situation, not a vague description:

I am a [year/level] student in [field] at [university].
My current courses: [list of main subjects].
I write in academic French.
My English level: [level] (for English-language sources).
I am preparing for [exam/competition/thesis] by [date].

The second field describes how the model should respond to you. This is where you eliminate behaviors that annoy you—summaries that are too short, forced English, or unsupported assertions:

- Respond in French unless I explicitly ask for English
- Use an academic but accessible tone
- Always structure responses with headings and lists
- Cite sources when possible (author, year)
- Warn me when information is uncertain
- Do not over-summarize: I prefer deep understanding

Spend ten minutes filling out these two fields today. You will see the payoff in every subsequent conversation.

Organize Your Conversations

The ChatGPT sidebar quickly becomes an unreadable pile of unnamed discussions, and you end up asking a question you already received an excellent answer for. Adopt a three-category naming convention from day one: [COURSE] Topic for subject tracking, such as [LAW] Civil Liability; [PROJECT] Name for long-term work, like [THESIS] Literature Review; and [REVIEW] Subject for practice sessions, for example [REVIEW] L2 Statistics. Rename your last five conversations using this scheme, and you will immediately see the difference when searching.

For work that spans weeks, Shared Projects—available with Plus and Edu plans—go further. They create a persistent workspace with shared files and context. For a thesis or dissertation, you upload your outline, sources, and drafts, so ChatGPT stops rediscovering your topic with every exchange. For a group project, the same space shares context with your teammates. For a semester of review, it keeps track of your progress. Create one via Menu → Projects → New Project, give it a clear name, and upload your reference documents.

Select the Model Before Asking

Not all models suit every task, and this choice often matters more than clever prompt engineering. Reserve GPT-5.6 Terra for quick questions, rephrasing, and short translations. Switch to GPT-5.6 Sol for long-form writing, text analysis, and outlining. Enable medium or high reasoning whenever the task involves math, logic, coding, or step-by-step proofs. Finally, Deep Research is the tool for in-depth literature research, which Lesson 4 details. Make it a habit to select the model before writing your question, and test the same prompt on Terra then on Sol. The difference in depth will serve as a lasting benchmark.

Useful Extensions and Integrations

On a browser, the web interface remains your primary workstation, offering access to all models. The browser extension lets you query ChatGPT from any page without switching tabs. The iOS and Android apps complete the setup with voice dictation: asking a question orally on the subway turns a commute into a review session. Regarding existing tools, you can paste texts from Google Docs or Word directly into a conversation for proofreading, and references exported from Zotero or Mendeley as text become usable by the model.

Key Takeaways

• Custom Instructions prevent you from repeating your context in every conversation
• Organize conversations with a consistent naming convention
• Shared Projects are essential for long-term work (theses, dissertations)
• Choose the model suited to the task before asking your question
• The mobile app is a powerful asset for reviewing on the go