← All articles
Education and Research

Presentation for ai

Narrow your AI presentation to one question, check sources yourself, and verify any AI-generated draft before using it. Includes a hypothetical example and a final checklist.

Draw Editorial Team7 min read

Define the purpose and scope

When preparing a presentation about AI, first narrow a broad topic such as “What is AI?” to a single question your audience wants answered. Decide on the presentation length, target audience, and assignment or event requirements from the outset, too. These will help you choose what information to include and how much detail to cover.

For example, if you are preparing a five-minute presentation for high school students, you could narrow the topic to “How is an AI system used for a particular task, and what questions should be answered in this example?” If you are introducing a research paper, examine its purpose, the study’s main contributions, and its examples. The course guidelines at Rose-Hulman recommend covering the purpose, contributions, any software demonstration, examples, and a brief assessment of why the contribution matters in a research presentation. The guidelines also ask you to discuss how the work could be extended: Course guidelines for an AI presentation. These guidelines are for a specific course; check the time limit and submission requirements for your own assignment separately.

Check sources and claims

Before drafting, gather your class notes, your teacher’s instructions, and sources related to the topic. For each claim, note the source’s name, link, and the section where the information appears. If you are drawing on a research paper, do not rely only on its abstract: examine the relevant text, figures, and tables. Do not include a visual in your presentation as a definitive finding without checking what it shows against the source.

Asking an AI tool for ideas, a title, or an initial draft does not replace reviewing sources. If the tool provides a source or technical explanation, open the link and check whether the information is actually there. If you cannot find a source supporting a claim, remove it or rewrite it as a question that needs verification.

Build the slide sequence

For a short presentation, you could try this sequence: title and guiding question; basic explanation of the topic; an example that makes the explanation concrete; limitations or open questions about the example; conclusion and sources. This is a draft structure you can use; the length of each section and the number of slides depend on the presentation time and the instructions.

Aim to have one main idea on each slide. If you use a technical term, check it against a reliable source and explain it in a way that suits your audience. When you add a visual or diagram, state what it shows and where you got it. The Rose-Hulman guidelines also recommend using diagrams and other visual aids from web pages or articles to make the work easier to understand. You should assess whether the visual really contributes to the point you want to make.

Example: “How might a system classify fruit?”

Hypothetical presentation plan: Imagine you are preparing a four-minute, five-slide presentation for middle school students. This example is not a verified research finding; it is a presentation outline to be filled in with sources.

  • Slide 1 — Question: “What information might a computer look at to group examples of fruit?” Say that you will answer this question during the presentation.
  • Slide 2 — Concept: Choose the technical term you will use. After checking its definition and its relationship to AI in a source directly relevant to the topic, explain it in your own words.
  • Slide 3 — Hypothetical example: Use a drawing showing possible features such as color or weight. Make clear that these do not prove how the system actually works and were chosen only for the example. If you explain how these features would be used, verify your explanation against a relevant source.
  • Slide 4 — Points to check: Ask questions such as “When might this example be misleading?” and “Is the information being used sufficient?” Do not present answers that are not supported by sources as definitive claims about real systems.
  • Slide 5 — Conclusion: Answer the original question based on the information you verified; distinguish the hypothetical example from a real finding and include your sources.

This plan makes the presentation’s main question and verification steps clear without portraying the example as a technical fact. If you are presenting a project in robotics, the connection between the project question and example of the work in the guide to preparing a robotics project presentation may be useful. You can also consult Technology presentation: simplifying a technical topic for ideas on simplifying a technical subject.

Ask an AI tool for a draft

The University of Virginia Library’s exercise suggests asking a tool for a draft of presentation titles, bullet points, and speaker notes, then comparing the result and revising it in your own voice. The output described in the exercise is not a ready-to-use presentation, but content suggestions to work on: UVA Library: presentation preparation exercise.

For example, you could start with this prompt: “Suggest a five-slide outline for middle school students focused on the question ‘How might a system classify fruit?’ Add a title, brief bullet points, and speaker notes to each slide. Label the fruit example as hypothetical; identify technical statements that need verification and places that require sources.” This prompt does not guarantee a correct answer or sources. Compare the output with your sources, remove inaccurate or unsupported claims, and check your class’s rules on AI use.

Review the draft and rehearse

If the draft is too general, rewrite the main question and consider how each slide contributes to answering it. If a slide has too much text, separate out the main message; if it suits your presentation format, move the details into your speaker notes. If a technical explanation is unclear, check the term and example again; if you use an analogy, point out where it may fall short.

Before the presentation, try explaining the main idea without reading the slide text. If you cannot explain a finding or term, return to the relevant source; if you cannot verify the information, phrase it more cautiously or remove it. If research presentations also require a short written summary alongside the slides, check that the summary and slides reflect the same purpose and key information. The Rose-Hulman guidelines specifically state that the summary and slides should be consistent with each other for their course.

Final checklist

  • Are the presentation’s audience, purpose, duration, and instructions clear?
  • Has the broad topic been narrowed to one answerable question?
  • Have technical explanations been checked against relevant sources?
  • Are hypothetical examples clearly distinguished from real findings or measurements?
  • Does each slide contribute to the main question, and are the sources of visuals identified?
  • Does the presentation fit the time limit, and can you explain the key points in your own words?

Frequently asked questions

Can I have AI prepare the entire presentation?

You can ask for a draft, a title, or suggestions for an explanation. But do not treat the tool’s output as verified content: check the claims and sources yourself, and follow your class’s rules on use.

How many slides should a presentation have?

There is no single number that works for every presentation. First check the time limit and instructions; then choose the slides needed to explain the main question.

← Back to blog