Now and again, I’ll get a screenshot that has some text on it that I’d like to use. In the old days, I’d have had to type it out manually - BORING!!! But now that AI is upon us, it can be done really quickly and easily. The only downside is that on a free ChatGPT account, there is a limit on how many files you can do per day.
All you need to do is:
Enter a prompt like:
“can you read the text on an image and reproduce the text?”
The AI will answer and ask you to upload the image.Upload the image and voila! The AI gives you the text from the image.
Copy and paste the text into whatever app you need it for. Edit the text if necessary.
Examples of how I’ve used AI transcription
Here are some examples of how I’ve used AI transcription on projects:
Improving UI text and descriptions
Sometimes, the developers I work with enter the most confusing, jargon-riddled, or just plain overlong feature names and descriptions. When I’ve figured out what the feature actually does, I’ll suggest improvements (UX writing!). For that, I like to include the original text and then my version. I take a quick screenshot, upload it to Chat GPT and ask it to give me the text. Copy and paste into the JIRA ticket and it’s job done.Alt text for images
This is a less common use, but sometimes I’ll have images that contain text and that text would be useful in the alt text description. For example, I’ve used AI to transcribe social media comments in screenshots and to get the text from callouts on diagrams.Getting text from lost documents
Let’s say I have a screenshot of a document in a knowledge base article. I need to update the screenshot, but I’ve lost the original document. This is where AI can make it easier to recreate the document - I ask it to transcribe the screenshot, then copy and paste the text into a new document. Hey presto! I have quickly remade the document and can take a new screenshot of it. And actually save the document somewhere safe this time.
