Generative AI and image editing tools answer two different questions. Generative tools — Pose AI, Midjourney — answer "what new image can you make me?" starting from a description (and, for identity-locked tools like Pose, a selfie). Editing tools — Photoshop with Firefly, RunwayML — answer "how do I change this image or video I already have?" The confusion happens because both categories now use generative AI models under the hood, so the underlying tech doesn't tell you which category a tool is in — what it starts from does.
This lays out the distinction plainly, since picking the wrong category for the job is the most common source of frustration with either kind of tool.
- Generative AI (Pose AI, Midjourney) creates a new image from a text prompt — Pose additionally locks the result to your identity from one uploaded selfie, where Midjourney has no identity input and generates a generic figure or scene.
- Image editing tools (Photoshop with Firefly, RunwayML) start from an image or video you already have and modify part of it, keeping the rest intact.
- For changing a pose specifically: if you have a photo and want the pose adjusted within it, that's editing (Adobe Firefly's Generative Fill). If you want a new photo of yourself in a different pose, that's generation (Pose AI).
- Neither category is "better" in general — they're suited to different starting points: a blank canvas plus a description, versus an existing image plus a change.
Key terms
Generative AI, in the image sense, means creating a new image from a text description (and sometimes a reference image or, for identity-locked tools, an uploaded selfie) — there's no existing photo being altered, the output is built from scratch. Pose AI's Image Studio (Nano Banana 2) and Midjourney are both generative in this sense, though Pose's output is locked to your actual identity from one selfie, while Midjourney generates images with no identity constraint.
Image editing, in this context, means modifying a specific image or video you already have — removing an object, adjusting lighting, changing a pose, extending a background — while the rest of the original stays intact. Photoshop's Generative Fill (powered by Adobe Firefly) and RunwayML's Edit Studio both fall in this category: they take an existing asset as input and change part of it.
Generative AI vs image editing tools
| Generative AI (Pose AI, Midjourney) | Image editing (Photoshop/Firefly, RunwayML) | |
|---|---|---|
| Starting point | A text description, plus a selfie for identity-locked tools like Pose | An existing photo or video you already have |
| What changes | Everything — the whole image is built new | Only the part you select or describe — the rest stays as it was |
| Identity preservation | Pose: yes, from one selfie. Midjourney: no, no identity input | Preserved by definition — it's the same original subject, just modified |
| Best for | A new photo that doesn't exist yet — a headshot, a themed portrait, a full-body shot | Adjusting something specific in a photo you already took or have |
Generate a new identity-locked photo from one selfie on the AI headshots page.
Turn a generated photo into a clip with Pose AI's Video Studio.