Please read the whole policy carefully. It applies to every assignment in CSE 541, and a course-specific addition follows the Examples section.
Overview
AI is welcomed in this course for collaboration in learning, problem-solving, and creating. You may use AI tools for any course work—including assignments, projects, quizzes, and exams—unless a specific activity states otherwise. Use AI actively and thoughtfully: its role is to help you explore, build, test, revise, and deepen your understanding. Read the full policy before using AI so you understand the privacy expectations and any activity-specific restrictions.
Definition of AI
For this policy, “AI” refers to generative AI systems that can produce or modify text, code, images, audio, video, or other content — whether in a single response to a prompt or by operating autonomously across multiple steps (e.g., chatbots such as ChatGPT, Claude, and Gemini; code assistants and agents such as GitHub Copilot, Cursor, and Claude Code; and AI-powered writing, translation, and image-generation tools). This includes AI features built into IDEs, operating systems, or browsers when they generate or substantially rewrite content, but excludes non-generative tools such as basic spell-check, grammar/syntax highlighting, and linters. A tool’s ability to act autonomously (e.g., independently writing, running, and debugging code) does not exempt it from this definition.
Rationale
This course treats AI as a full collaborator in the learning process: it can be used to ask questions, explore possibilities, build and debug software, and refine your work. AI can accelerate experimentation and make feedback more available, but you remain responsible for directing the collaboration, evaluating its output, and developing the understanding behind the work you submit.
Permitted actions
- You may use AI tools for any course-related task, including brainstorming, concept explanations, summarization, drafting, coding support, debugging assistance, editing and revision, preparing for quizzes and exams, and completing assignments and projects.
- You are encouraged to use AI collaboratively: ask it to explain its reasoning, propose alternatives, challenge your assumptions, review your work, and help you investigate errors.
- If an activity includes additional restrictions, those restrictions will be stated explicitly.
Prohibited actions
- Do not use AI in ways that violate any activity-specific restrictions stated by the instructor.
- Do not represent AI-generated content as independently produced.
- Do not violate data privacy rules.
Examples
- Allowed: collaborating with AI to brainstorm project ideas, explain a concept, generate an outline, propose test cases, or compare solution approaches.
- Allowed: using AI to help write, debug, document, or improve code for an assignment when no activity-specific restriction applies.
- Allowed: using AI to critique your reasoning or code, then using that feedback to revise your work and strengthen your understanding.
- Allowed: using AI to prepare for a quiz or exam when that activity permits AI use.
- Not allowed: using AI on an activity that explicitly prohibits or limits it.
- Not allowed: uploading restricted course materials or sensitive information to an AI tool.
CSE 541 addition: you supply the algorithm and the proof
This narrows the “Allowed: using AI to help write … code for an assignment” example above. On homeworks and projects, you may use AI to turn an algorithm you have fully specified into code, but you may not ask AI to supply the algorithm. For example, prompting “Implement UCB” is not acceptable. Writing out the UCB index, the arm-selection rule, the observation model, and the loop, and then asking AI to code it up, is acceptable. See the full side-by-side example in the collaboration policy.
For proof-based questions, use AI minimally: to clarify a definition or concept, brainstorm possible approaches, or help with LaTeX formatting. Do not ask AI to write, complete, or fix your proof.
If you’ve read this far, thank you. Complete this short quiz (Google Form, opens in a new tab) by October 11. If your final grade ends up right at a cutoff, it will be bumped up.
Required attribution
- If you used AI on a graded assignment, include a short note specifying the tool you used and how you used it.
- Example: “I used Claude Code to develop and debug the code for this assignment.”
Guidelines for effective AI use
- Treat AI as an active collaborator: give it context, ask follow-up questions, and use its responses to extend rather than replace your own thinking.
- Check important output, especially code, factual claims, citations, and calculations; AI can be confidently wrong.
- Use AI to explore alternatives, find mistakes, generate tests, and iterate on your work.
- Be prepared to explain and defend work you submit, including choices you made with AI assistance.
- Follow all assignment-specific directions about permitted tools, disclosure, and attribution.
Data privacy expectations
- Do not upload course-specific materials—including assignment prompts, course code, datasets, class notes, hidden tests, solution keys, or other restricted content—to AI tools without the instructor’s explicit approval.
- Some AI tools (e.g., Cursor, Claude Code, and similar coding agents) can automatically access files beyond what you explicitly share — including your entire local repository, file system, or connected accounts. Before using such a tool, confirm what it can access and restrict or scope that access so it does not read private repositories, hidden test files, other students’ code, solution keys, or other restricted materials without your intent.
- Do not submit personal, sensitive, or university-protected information into AI tools.
- Be aware that prompts and outputs may be stored or reused by third-party AI providers.
- UW’s Purple tool and certain GenAI tools provided by the Allen School for use in courses are designed specifically for UW students and staff and have stronger privacy protections than public commercial tools. When AI usage is permitted, using Purple or Allen School-provided tools is advised if possible.
Academic integrity
Using AI in ways prohibited by this policy or by assignment-specific instructions will be considered academic misconduct and referred to CSSC. Students found responsible for violating this policy will, at minimum, receive no credit for the affected assignment. Stronger penalties, up to and including failing the course, may apply for repeated or egregious violations. In cases of group work, all students will be held equally responsible for violation unless there is clear evidence that a subset of team members were not involved or aware.
Accommodations and approved exceptions
Students with accessibility needs or other extenuating circumstances should contact the instructor to discuss exceptions. Accommodations or exceptions require instructor approval before they are used.
Policy text: Allen School Course AI Policy Templates (Autumn 2026 edition), Policy 5, developed by the Allen School Teaching with Generative AI committee. Questions about how it applies to a specific assignment? Ask on Ed.