怎么用 AI 做 PPT[object Object]
微型指南 · 约 8 分钟Guide · about 8 min · 引用核验 2026-11-19(仅引用链接与日期) · refs checked 2026-11-19 (reference links and dates only)
让 AI 先做结构、再做文案、最后做视觉,避免一上来就生成一堆套模板的空话。Let AI do structure first, copy second and visuals last — not a stack of template slides with generic claims.
先给受众、场景、时长、要做的决定;这四项决定页数和每页信息密度。
让模型先出「每页一句话结论 + 证据」的大纲,人确认后再扩成正文。
每页只放一个结论,图表服务结论;不要把整段报告拆成 30 页 bullet。
数字、引用、竞品名逐条回到原始来源核对,AI 生成的数字默认不可信。
视觉统一用品牌色和两档字号;生成后检查中文字体、换行和图表标签是否被截断。
最后让模型当评审:删掉没有结论的页、合并重复页、给每页写一句讲稿。
Give audience, setting, duration and the decision to be made; those four set page count and density.
Ask for a one-sentence conclusion plus evidence per page; approve the outline before expanding.
One conclusion per page; charts serve the conclusion, not a report chopped into 30 bullets.
Verify every number, citation and competitor name against the original source; AI-generated figures are untrusted by default.
Use brand colours and two type sizes; check Chinese fonts, line breaks and clipped chart labels.
Have the model review: delete pages without a conclusion, merge duplicates, write one speaker line per page.
例子与判断Example and interpretation
以下是教学情境,不是本站运行或实测记录。A teaching scenario, not a measured or executed site test.
先让模型给 10 页大纲,每页一句结论;你删掉两页、改三页结论,再让它扩成正文和讲稿,最后核对所有数字。Ask for a 10-page outline with one conclusion per page; delete two pages, rewrite three conclusions, then expand into body copy and speaker notes, and verify every number.
做完例子后怎么检查How to check your practice
先用自己的话解释上面的判断,再检查输入条件,最后回原文验证。说不清原因时回概念;缺日期、字段或权限时补输入;结论无出处时先撤回结论。Explain the example in your own words, inspect input constraints, then verify against the source. Revisit concepts if you cannot explain why; supply missing dates, fields or permissions; withdraw unsupported conclusions.
按这篇主题,先从“概念没分清”练习查起For this topic, start with concept confusion · 查看全部六个情境与理由See all six scenarios and their reasoning
结构化输出Structured Output 文生图Text-to-Image Generation 人工复核队列Human Review Queue