A Continuum of Uses of AI

When you shift your perspective from thinking about AI as an independent agent that “creates” things to seeing it as a tool that can expand your capacities, AI use becomes less of a binary action and more of a continuum of ways it can assist you.  As you consider the levels of AI use below, consider how your goals impact your perspective on the roles AI might play in the process.

When humanizing AI, e.g., as an “AI Tutor”, “AI Assistant” or “AI Grader”, be cautious of making false equivalencies with and devaluing/dehumanizing the valuable work that humans in these roles produce.

Planning & Organizing 

(AI helps you structure your work, but you remain in full control.)

AI can help you plan your approach to projects and navigate challenges along the way.

  • Breaking down complex tasks → Suggest ways to break up a project into parts or structure a research plan.
  • Tracking and scheduling → Map out deadlines and create a workflow.
  • Overcoming procrastination → Provide encouragement, time management strategies, or check-ins to keep you accountable.
  • Interactive encouragement  → A thinking partner to reflect on ideas and refine next steps.
Example: You’re struggling to start a final project for a course. You open up a new chat window and start telling the AI system about the project and your initial thoughts and ideas.  After uploading the project instructions and assignment, you ask AI to suggest a timeline with key milestones and potential barriers you might encounter. As you make progress, you use AI as a space to chat about your ideas, asking it to prompt you in your next steps and revise the timeline as needed.

Reading & Understanding 

(AI helps you comprehend and engage with materials, but you directly engage with the text to critically assess its responses.)

AI is both useful and flawed in summarizing complex texts. When you feed most systems a document (or a meeting transcript) and ask them for a summary, you typically get back a clear, succinct set of ideas.  Unfortunately, generative AI is not pulling on a conceptual map of the key ideas in the text, it is instead generating its summary based on the context you provide it with in your prompts, the words used in the text, and its training dataset. Consider using AI as a reading companion, especially if you are having trouble focusing on the reading, or encountering lots of unfamiliar concepts.

  • Clarifications & Definitions → Explain confusing or new concepts in simpler terms, create analogies or examples.
  • Summarizing → Generate concise summaries of a text, compare and contrast key themes across multiple texts to help focus your reading. Some AI’s, e.g. notebook.google.lm, are designed to help you collect and keep notes across different text sources.
  • Engaging with material → Provide feedback about your developing ideas related to the text, generate discussion questions.

Caption: A snapshot of Ask AI Assistant popping up in Adobe’s Acrobat Reader.

Activity: Try summarizing a reading yourself, then compare it with AI’s version—what does it get right or miss? Tests like this, especially considering different kinds of content, can be important to understand when encountering ‘AI assistants’ in programs like this snapshot of an “Ask AI Assistant” from Adobe.

Exploring & Generating Ideas 

(AI helps expand your thinking, but you verify and refine its suggestions.)

Generative AI can be a powerful brainstorming partner to expand your ideas —but how you use it makes a difference.  Biases in the training data and the context you provide in your prompts can lead the AI to overlook critical perspectives and overemphasize certain directions.  Remember AI fabricates sources, so always fact-check the outputs.

  • Connect ideas → AI can suggest links between topics such as connecting your interests to an area you are studying.
  • Expand research directions → AI can point you toward journals, authors, or research areas worth exploring. Consider using specific AI’s built to help explore academic literature (e.g., Research Rabbit, Elicit, Scite.ai, Consensus, etc.)
  • Challenge your thinking → Explore counter arguments, missing concepts or alternative perspectives.
Example: You’re brainstorming research questions for a project and share your ideas and the project constraints with an AI tool. The tool can suggest interdisciplinary connections between topics you are interested in, note potential important ideas or topics you haven’t yet considered.

Technical Assistance & Problem-Solving 

(AI helps troubleshoot tasks, but you evaluate and apply solutions.)

AI can be a useful tool for practical problem-solving, especially in technical fields, where solutions can be tested in real time.

  • Finding errors in code or calculations → AI can debug code, suggest solutions, or explain why something isn’t working.
  • Learning software tools → AI can guide you in using new programs, from statistical analysis software to citation managers.
  • Generating multiple approaches → Ask AI for different ways to approach a problem and decide which is best.
  • LMGTFY  → Describe the challenge you are encountering and what you’re trying to accomplish in your own words and AI can help you find the language you need to better search the web for solutions.
Example: You’re analyzing data and your statistical software keeps returning an error. You ask the AI to help you decode the error messages, and suggest multiple solutions for you to test out and find what works.

Data Analysis –

Chat GPT is happy to make yo9u a plot, reformat your data, etc. but it’s generative not algorithmic – so BAD idea. Chat GPT is GReat at running the sttat software package which then maintains fidelity in the data anlysis …

example image from AI FOrum …

Writing Support & Editing 

(AI refines your writing but does not replace your ideas or voice.)

AI can excel at editing tasks and suggest ways to improve written work. However, without careful prompting the feedback it gives can be superficial, and if it rewrites your text for you – it can insert new meanings that are difficult to detect in a quick readthrough. Therefore, in this stage of the continuum you don’t ask AI to rewrite your text for you (don’t copy & paste!) but to give you ideas that you can alter the text with.

  • Outlining & Structuring → Help organize notes into a potential outline.
  • Feedback & Revision → Suggest improvements, clarify arguments, or offer alternative perspectives.
  • Grammar & Style Checks → Highlight typos, suggest clearer phrasing, or provide suggestions to refine writing for specific audiences.
Microsoft’s Copilot writing assistant in Microsoft Word 365 at Tufts
Example: You’re writing a paper to communicate new research findings. You ask AI to develop some potential counterarguments for your thesis.  You address a few that you find most relevant in your paper.  

Co-Writing

(AI drafts content, but you substantially rewrite and refine it.)

At this level, AI is not just assisting—it’s producing content that you must actively shape, revise, and make your own.

  • First Draft Assistance → Generate draft text based on your outline, which you then rewrite extensively.
  • Style Adjustments → AI can help rewrite a section in a different tone or format.
  • Checking AI-generated content for accuracy → AI can fabricate facts, so critical review is necessary.
Example: You provide AI with an outline and ask it to draft a section of your paper, then rewrite it entirely in your own words.

Cautions: 

  • Veracity: Experts on a topic will often find more fault with the generated text, finding answers to be surface level, contain misleading or even fully incorrect information.
  • Voice: The default voice of AI is not as engaging as having ideas expressed through your own voice as a unique author.  While there are places, eg. lab reports, legal briefs, etc. where the voice of the author is not a key component, in many cases it is more engaging for the reader to hear the varied voices that come from different authors.
  • Citations: Using AI generated text without citing or attribution can violate policies (e.g., course syllabus, journal citations) and at times be considered plagiarism.  Carefully consider how you will indicate your use of AI when using it.
  • Copyright: https://www.copyright.gov/ai/