# AI Research Seminar Slides From Your Own Notes

Use AI to explain work you have actually done, not to manufacture findings. Prepare a source bundle and a clear research question, then draft a slide structure that separates prior work, your method, observations and limitations. Follow your institution’s rules on AI assistance and attribution.

## Your brief

1. Provide your research question, approved abstract, method notes and actual findings. If results are pending, say so and frame the talk as work in progress.

   ____________________

2. Attach or cite only material you may share. Preserve source identifiers, page numbers, table units and the distinction between a quotation and a paraphrase.

   ____________________

3. Specify audience knowledge, speaking time and assessment requirements. Identify which parts need advisor or co-author approval before public presentation.

   ____________________

## Prompt

Build a seminar slide outline from [my notes and sources] for [audience] in [minutes]. Include question, background, method, findings, limitations and discussion. Cite supplied sources in notes. Do not invent experiments, statistics or references. Mark unsupported claims and missing results. Suggest visuals only where the data supports them and list questions I must answer before presenting.

## Illustrative example

Illustrative notes contain five interviews from an exploratory project. A responsible slide describes recurring themes and the small sample, with anonymized excerpts authorized for use. It does not convert those five interviews into a population-level percentage or claim statistical significance. If a chart is proposed, its denominator and coding method must be stated.

## Acceptance checks

- [ ] Can each finding and citation be traced to an actual supplied source, with units and denominators preserved?
- [ ] Are sample limits, unresolved questions and work-in-progress states visible?
- [ ] Does the exported deck keep citations readable and comply with the course or research group’s disclosure rules?

Do not ask AI to ‘add references’ without providing verifiable sources. A plausible-looking paper title or citation is not evidence. Open references, confirm their relevance and get permission for unpublished or identifiable material.

---

Brainbaby · https://brainbaby.ai/blog/ai-research-seminar-slides

CC BY 4.0 — https://creativecommons.org/licenses/by/4.0/