As of 13 August 2026, AI can only partly compress your photos without losing quality.
This still needs a person who signs their name to it.
Can you do it?
5 minutesto a draft.
30 minutesto something you’d act on.
Cost, all in£0
Skill neededchat-fluent
Who has to check ityou
What the alternative costsTopaz Labs is a professional AI tool for photo and video upscaling, denoising and sharpening.
If this goes wrong: you upload or share a photo that looks softer or has lost useful information, while the original remains available if you keep it.
What to actually do
Hand it to a person
The route this page recommends
A person who owns the outcome does this end to end, worth it when the failure is dear.
Do it yourself
Second choiceA chat interface, chat-fluent skill, and roughly 30 minutes until you can act on the result.
How to actually do it
- Make a backup of the original photos and work only on duplicate copies so the source files are not overwritten.
- Open the file details for a representative photo from each batch and record its format, pixel dimensions, file size and whether it contains metadata you need to keep.
- Record where the copies will go, such as an email attachment, website, online form or cloud folder, and note any stated file-size or format limit.
- Paste the prompt into a chatbot with those details and ask it to distinguish lossless compression from a visually acceptable smaller export.
- Open an image editor or export tool such as Fotor, import the duplicate copies, and apply only the format, dimension and quality changes that match the destination and the chatbot's advice.
- Export the results as new files, then compare each result with its original for dimensions, file size, metadata and visible detail at 100 percent, including faces, text, edges and fine texture.
- Keep the compressed copies only if they meet the destination's limit and still look right for their intended use; retain the originals for printing, editing or future exports.
Prompt
I need to make copies of photos smaller for [upload, email, website or storage]. I want the smallest files that still look indistinguishable from the originals at normal viewing size, but do not claim that compression is lossless unless it genuinely is. For each photo or batch, use these details: format [JPEG, PNG, HEIC, RAW or unknown], current dimensions [width x height], current file size [size], intended destination [destination], maximum file size if any [limit], and whether metadata must be kept [yes or no]. Tell me whether lossless compression is realistic for this format and use, whether changing dimensions is necessary, and which export format and quality approach to try. Give a short step-by-step workflow for making copies, not overwriting originals. Include a checking method using file size, dimensions, metadata and a visual comparison at 100 percent. Do not invent settings supported by a particular app, and state what I need to confirm in the app before exporting.
Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.
Use a tool built for this
The distant third
What it gets wrong
- AI cannot make a lossy JPEG or similar file smaller while guaranteeing that no information or visible quality has changed.
- It cannot know whether your destination will reject a format, strip metadata or impose a hidden size limit unless you provide that information.
- It cannot reliably judge every subtle change in colour, texture, sharpening or fine detail from a setting recommendation alone.
- It cannot restore detail that has already been discarded by an earlier export.
- It cannot safely choose between a smaller file and higher quality without knowing how you intend to view, print or share the photo.
What caps this at PARTLY: verification cost, judgement under ambiguity and stakes of error.
How we scored this
Five axes, each scored nought to two by hand: ten means AI carries the task cleanly, and the thresholds that turn a total into YES, PARTLY or NO are published in the methodology. Each axis name links to its definition.
| Axis | Score (0–2) |
|---|---|
| Output | 1 |
| Inputs | 2 |
| Verification | 1 |
| Liability | 2 |
| Effort delta | 1 |
| Total | 7 / 10 |
The methodology and its thresholds are published in full.
FAQ
- Can AI compress photos without losing quality?
- Partly. AI can recommend lossless or visually acceptable settings, but making a file smaller often involves discarding some data, and only you can decide whether the change is acceptable for the intended use.
- What is the best format for compressing photos?
- It depends on the original format and destination. Ask AI to compare the formats your destination accepts, then test an exported copy while keeping the original unchanged.
- Will compressing a photo make it blurry?
- It can, particularly when quality is reduced or the pixel dimensions are changed too far. Compare the compressed copy with the original at 100 percent and inspect faces, text, edges and fine detail before sharing it.
- How do I reduce photo file size without losing quality?
- Start with a duplicate, identify the destination's format and size requirements, and try lossless compression where the format supports it. If that is not enough, export a visually acceptable copy and compare it with the original before sending or uploading.
Nearby answers
Assessed by gpt-5.6-luna (gpt-5.6-luna) on 2026-08-13, second-checked by an independent model. Wrong somewhere? Email [email protected] and it gets re-checked.
The newsletter
AI news, new answers and product picks, straight to your inbox.