Home · Personal · Writing & Creative · Photo & video
As of 13 August 2026, AI can translate a sign in a photo.
Most people should do this themselves; the prompt is on this page.
Can you do it?
2 minutesto a draft.
15 minutesto something you’d act on.
Cost, all in£0
Skill needednone
Who has to check ityou
What the alternative costsThe supplied tool list gives no price for a human translation alternative.
If this goes wrong: the model misreads a word or misses context, so you get an incorrect translation and need to confirm it before relying on the sign.
What to actually do
Do it yourself
The route this page recommends
A chat interface, no skill needed, and roughly 15 minutes until you can act on the result.
How to actually do it
- Open a chatbot that accepts image attachments, such as ChatGPT, Claude or Gemini.
- Take a new photo of the whole sign in good light, keeping the camera straight and close enough for every word to be readable.
- Attach the photo and paste the prompt above, adding the country or likely language if you know it.
- Read the model's transcription against each visible word, number, symbol and line break in the photo.
- Ask for a second translation if a word is marked uncertain, the sign is partly obscured, or the wording could be a warning or restriction.
- Before acting on an important instruction, send the photo and both translations to a fluent speaker or ask the sign's owner to confirm the meaning.
Prompt
Translate the sign in the attached photo into plain British English. First transcribe the original text exactly as far as you can, preserving line breaks where they help. Then provide the translation. Identify the likely source language, mark any unreadable or uncertain words, explain ambiguous phrases or symbols, and do not guess silently. If the sign contains a warning, prohibition, safety instruction or legal restriction, quote that part separately and tell me to confirm it with a fluent speaker before relying on it.
Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.
Use a tool built for this
Second choiceHand it to a person
The distant thirdA person who owns the outcome does this end to end, worth it when the failure is dear.
What it gets wrong
- It cannot reliably recover words hidden by glare, blur, damage or an unusual font.
- It cannot know local slang, missing background or the sign's intended legal meaning from the image alone.
- It can produce a smooth translation while misreading one short word, number or symbol.
- It cannot take responsibility for what you do after following the translation.
Even on a YES, the friction has a name: stakes of error, context depth and judgement under ambiguity.
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 | 2 |
| Inputs | 2 |
| Verification | 1 |
| Liability | 1 |
| Effort delta | 2 |
| Total | 8 / 10 |
The methodology and its thresholds are published in full.
FAQ
- Can ChatGPT translate text from a photo?
- Yes. Upload a clear photo and ask it to transcribe the original text before translating it, so you can compare the transcription with the sign.
- How accurate is AI translation from a photo?
- It is often useful for clear, printed signs, but accuracy falls with blur, glare, handwriting, unusual scripts and missing context. Check every uncertain word and confirm warnings or restrictions with a fluent speaker.
- Can AI translate a sign into English?
- Yes. Ask for plain British English and request the likely source language, an exact transcription and a separate explanation of any ambiguous wording.
- Is it safe to rely on AI to translate a warning sign?
- Not without confirmation. A missed word or symbol can change the instruction, so have a fluent speaker or the sign's owner check it before you act.
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.