AI Basics
AI Basics
Get started with AI use cases and tips for basic prompts
This page includes a high-level summary of some basic use cases for AI tools, and a guide to help you get started with prompts.
- Before using AI tools at Harvard, be sure to review the University’s guidelines and use an approved tool.
- Find more guidance on advanced AI features such as agents, assistants, coding, and connectors.
- Learn about the many ways that the Harvard community is using AI.
Basic use cases
Content creation and review
AI can create or generate ideas for text, documents, images, presentations, music, videos, surveys and much more in seconds from simple, natural language prompts. It can also be a helpful partner in editing and enhancing your work, looking for errors, or suggesting areas of improvement.
💡 Keep in mind:
- AI-generated content, even when it looks professional or polished, can be inaccurate, generic, biased, or contain copyrighted information. Per University guidelines, review all AI-generated content before publishing or sharing.
Search
AI can quickly search from thousands of different sources from across the web. Natural language prompts make it easy to describe what you’re hoping to find even if you don’t know the exact terminology. Some tools include a “deep research” feature; an AI agent that works in the background to produce a thorough research report on the topic of your choosing, citing sources as it goes.
💡 Keep in mind:
- AI can return incorrect, outdated, or irrelevant information. Even when AI cites a source, that doesn’t guarantee that it has understood the source or summarized it accurately. Verify important information against original sources.
Data analysis
AI can analyze and generate concise summaries of long documents. This can be used for highlighting key insights from research papers, summarizing emails, reports, and meeting notes, or identifying key findings from spreadsheet data or surveys. Some tools have a “source-grounded research” or “notebook” feature that allows you to upload a select group of files or links and ask AI to analyze, summarize, or create content such as presentations from the data contained in your sources.
💡 Keep in mind:
- AI can overlook important context, misinterpret sources or data, make calculation errors, or draw conclusions that the source material doesn’t support. Double-check important findings and calculations against the original material.
Accessibility
AI can aid in making content more accessible by identifying problems in your documents, suggesting improvements such as alt-text and audio descriptions for images, or converting speech to text for captioning. Harvard’s Digital Accessibility Services can help you navigate the landscape.
💡 Keep in mind:
- AI-powered accessibility checkers can be a valuable time saver, but they can’t catch every error. You should still be sure to manually test your content for accessibility.
Get started with prompts
AI tools analyze your inputs (“prompts”) and generate a response based on the patterns it has learned through its training. Today’s AI tools are much better at understanding natural, conversational language than even a few years ago, so it’s less important to master the perfect prompt and more important to focus on your objectives and how the AI can help you get there.
Be specific
Generic prompts will produce generic results. Clear, specific prompts will help the AI understand what you want and produce better results. In your instructions, clearly state:
- What you want (your goals or desired tasks)
- e.g. “Produce a report…; analyze this data…; find information on…”
- Use examples
- What you don’t want
- Set any relevant boundaries or limits to help focus the output
- Relevant context
- Any background details on the purpose or broader context that might help the AI better understand the ask
- Tone, audience, style
- For whom and / or what purpose is this for, and how does that affect how you want it presented?
- e.g. Marketing brief, lesson plan, welcome speech, financial report
- Format of output
- e.g. Text document, spreadsheet, presentation, poster, annual report, etc.
Break it into chunks
You don’t have to get everything into your first prompt. Treat your prompts as a conversation with the AI. Start with a basic request and add to it over time. Change the wording or tone or add more context and specificity to guide the AI toward the output you’re looking for. Go back and add anything you think you’ve missed, or upload excerpts from a file instead of an entire document to keep the conversation focused.
Correct mistakes and give feedback
Chat with the AI as if it’s a colleague or teammate and you’re working on a project together. Give feedback – tell it which parts of the output were useful, and which parts could be improved. If you notice it got something wrong, tell it so it can correct its mistake.
Ask it for help
Stuck on how to create good prompts? Ask AI to help! Describe what you’re trying to accomplish, even if your idea isn’t fully formed, and ask the AI to help you determine the best approach. Ask “what information do you need from me to do this well?”