Claude
Key University AI Guidelines (Click to expand)
- No uploading copyrighted articles, book chapters, or subscription content.
- No uploading student work or personal data.
- AI hallucinations apply to all tools — verify everything.
- Use Library‑integrated AI tools for summarising academic articles.
- Use reference managers (EndNote, Mendeley, Zotero) — AI citations are unreliable.
- Gemini: connecting Gmail/Docs/Drive exposes personal data.
- Follow IMPS guidance for research and teaching.
Claude, developed by Anthropic, is renowned for its highly nuanced, natural writing style and its large "context window," meaning it can process and remember significant amounts of text in a single conversation without losing the thread.
- Best for: Nuanced, human-sounding writing and stylistic drafting.
- Good at: Matching specific brand tones and comparing multiple pieces of text.
- Safety note: Only upload content you have the rights to. Do not upload copyrighted academic articles or personal student data.
- Example prompt: "Read this previous newsletter I wrote to understand my tone. Now, write a new 300-word update about the campus sustainability initiative, matching that exact writing style."
Unique Features (Artifacts & Context)
Claude includes several unique user interface features that improve productivity for advanced tasks:
- Artifacts: When Claude generates code, web designs, or complex text documents, it displays them in a dedicated side-panel (an Artifact) so you can view, edit, and copy the work without scrolling through the chat history.
- Projects (Pro Feature): Allows you to group chats together and upload permanent background knowledge (like style guides or non-sensitive policy docs) that Claude will reference across all conversations in that Project.
- Large Context Window: Claude can read the equivalent of a medium-sized novel in a single prompt, making it unparalleled for deep textual analysis of large, uncopyrighted reports.
Document Comparison
Analyse themes across text. Claude is highly capable of taking multiple (non-copyrighted) documents and finding commonalities.
Prompt:
"Attached are three different internal departmental policy drafts. Compare how each handles flexible working arrangements and provide a summary table of the differences."
Tone & Style
Generate less robotic writing. Many users find Claude's output to be less formulaic than other models, making it ideal for drafting sensitive communications.
Prompt:
"Draft a welcoming email to new staff joining the department. Ensure the tone is warm and supportive, avoiding corporate jargon."
Fact-Checking and Hallucinations
Generative AI predicts the next logical word; it does not "know" facts. This means Claude can occasionally present false information or non-existent policies as absolute fact. Always verify specific statistics, dates, and claims against trusted university sources.
Data Privacy Warning
Unless you are using an explicitly procured enterprise version of Claude, do not upload documents containing sensitive personal information, financial data, or classified research. Public models may use inputs for training.
These pages are currently in beta. To give feedback, please contact DTS at ai.training@reading.ac.uk.