1L Guidance on AI Use

Written for incoming 1Ls. This guide explains how to use AI tools as study support without substituting them for legal reasoning.

First and most important: course-specific rules control. General Project guidance does not override a syllabus, assignment instruction, exam rule, clinic rule, journal rule, or supervisor instruction. When in doubt, ask the professor before using AI.

For current Penn Carey Law tool access and setup, use the AI Resources Portal.

The frame

Your main task as a 1L is to develop legal reasoning. AI can support that learning, but it cannot replace the work of reading, analyzing, writing, revising, and explaining legal arguments yourself.

Use AI as a study partner after you have done the first pass. Do not use it as a shortcut around the work that builds judgment.

How generative AI works

Claude, ChatGPT, Gemini, and similar systems are large language models. They generate likely text based on patterns in their training data and the prompts you provide. That can produce fluent and useful output, but it is not the same as legal reasoning, source retrieval, or truth checking.

The practical consequence: the answer may sound right even when it is wrong.

Where AI helps

AI is most useful for low-stakes study support:

• turning your notes into practice questions;
• asking you follow-up questions after you explain a rule;
• helping organize your own outline;
• generating hypotheticals after you have studied the doctrine;
• suggesting alternative ways to explain an idea;
• proofreading for grammar or clarity when permitted.

Where AI falls short

AI is weak where legal education most needs precision and judgment. It may invent cases, misstate holdings, flatten ambiguity, omit counterarguments, or give a confident answer where the right answer depends on context.

Do not rely on AI for case names, quotations, citations, holdings, procedural posture, statutory language, or professor-specific course takeaways unless you verify against the assigned materials or another authoritative source.

Critical risks

• Hallucination. AI may invent authority or distort real authority.
• Overreliance. If you use AI before you have done your own reasoning, you lose the practice law school is designed to give you.
• Bias. AI can reproduce biased assumptions from its training data. Check outputs carefully, especially in areas where race, gender, immigration status, disability, class, or policing affect legal outcomes.
• Integrity. Misrepresenting AI output as your own work can violate academic-integrity rules. Undisclosed AI use may also violate a course or assignment policy.
• Confidentiality. Do not paste another person’s work, confidential facts, client or clinic information, class recordings, exam materials, or anything you would not want treated as uploaded data unless your professor has expressly authorized that use and the tool is approved for that material.

The human-AI-human workflow

Use AI in three steps:

1. Human first. Read, brief, outline, or draft before asking AI for help.
2. AI second. Ask for questions, critique, alternatives, or organization.
3. Human last. Verify, revise, and decide what to use.

Only use AI for tasks where you have enough knowledge to evaluate the output. If you cannot tell whether the answer is good, the tool is not helping you yet.

Suggested uses

Good study uses include:

• “Ask me five questions about this case based on my notes.”
• “Give me a hypo that tests this rule, but do not answer it until I try.”
• “Compare these two doctrines in a table using only the notes I provide.”
• “Identify what is unclear in my explanation of this concept.”
• “Generate flashcards from my outline.”

Riskier uses include:

• “Brief this case for me” before you have read it.
• “Write my assignment” or “draft my answer.”
• “Find cases that support my argument” without checking the sources.
• “Tell me what my professor wants” when you have not supplied course materials.

Prompting as a skill

Good prompts give context and limits:

• Say the course and topic.
• Say what materials the tool may use.
• Say what you have already done.
• Ask for questions or critique before answers.
• Ask the model to identify uncertainty.
• Specify the format you want.

Example:

I am a 1L studying Contracts. I have already read the assigned case and written the rule in my own words. Ask me three questions that test whether I understand expectation damages. After each answer, tell me what I missed and what part of my notes I should review. Do not give me a full model answer unless I ask.

The big takeaway

AI can help you practice, organize, and test your understanding. It cannot replace the work of building legal judgment. Use it to enhance your learning, not to avoid the hard parts.

Status

Maintained for the Penn Carey Law community. Current tool-access details are in the AI Resources Portal.