Does the Translator API returns malignant html tags or scripts by any chance?

Myサイト１高電社 0

I have a question about the Microsoft Translator text API.

Are there any chances that the API returns translated words with malignant html tags or scripts, for any exceptional cases like the translation engine has happened to learn from some malignant input data?

For instance, when Japanese words that correspond with English “Corporate Information” has sent to the API, it returns the words “Corporate Information” with a malignant html code like '<a href="[a fishing website URL]">Corporate Information</a>'.

romungi-MSFT 45,961 Reputation points Microsoft Employee

2023-09-13T06:16:49.62+00:00

@Myサイト１高電社 Ideally, it should not add additional information with the response that is not part of the translation text. I would suggest posting your HTML input of the source language to check this further. Based, on my experience I have noticed previous translation requests respond with additional words or incorrect translation but not something related to malignant information. If this behavior is seen, it has to be reported to translation team to check what could be wrong with the model behind the API. Thanks!!

Share via

Does the Translator API returns malignant html tags or scripts by any chance?

Your answer