I have been researching an issue I am having with both the Samsung Keyboard and Google Voice Keyboard. Both of these keyboards are briefly displaying inappropriate content when I initiate the speech to text function. These inputs are very creepy, from hey baby to highschool girls ect. I do have a screenshot.
Have you heard of any other issues out there with text to speech having this type of vulnerability to hacking?
This began with the use of Quizlet AI for study, using AI for a verbal input question and answer is quite useful. The behavior spread from there to all of the text to all apps with text to speech inputs. Bing, Verizon texting, Chat GPT.
I did see a brief mention of curl being able to capture text to speech when working with AI, but can't find the article again.
Android Phone Galaxy S 21 One UI version 5.1 Android Version 13
I have factory reset my phone, but I am afraid to turn on the text to speech feature again to test this out.