Yesterday it came to light that there’s a bug on iOS devices where if you use the dictation function, and say the word “Trump” as in Donald Trump, it prints the word “Racist”. Now I did test this and I could reproduce this. But I could get other words to appear. But it really doesn’t matter at this point as the planet is hung up on the “Trump” and “Racist” connection. And all the usual nonsense that you would expect to happen is happening.
Scott Stephenson, Founder and CEO of Deepgram had this comment:
“AI is only as smart as the data it’s trained on. Voice recognition should be about understanding, not assuming. This is a reminder that companies need to constantly refine their models to avoid bizarre and potentially harmful mix-ups. The goal isn’t just accuracy, it’s trust.”
“Voice AI is exploding because, let’s face it, talking is faster than typing. But speed means nothing without precision. If AI can’t truly understand what we’re saying – accents, slang, emotions – then it’s just noise. The next wave of innovation is about making AI listen smarter, not just faster.”
On one hand, I am surprised that this wasn’t caught in QA. But on the other hand, I am not surprised because how far do you go to test a speech to text engine like this one? You can’t test every single word or name out there. Thus you can expect something like this to happen again. And chances are there’s nothing nefarious going on.
Like this:
Like Loading...
Related
This entry was posted on February 26, 2025 at 12:26 pm and is filed under Commentary with tags Apple. You can follow any responses to this entry through the RSS 2.0 feed.
You can leave a response, or trackback from your own site.
The World Is Losing Its Mind Over The Bug Related To Trump On Apple Devices
Yesterday it came to light that there’s a bug on iOS devices where if you use the dictation function, and say the word “Trump” as in Donald Trump, it prints the word “Racist”. Now I did test this and I could reproduce this. But I could get other words to appear. But it really doesn’t matter at this point as the planet is hung up on the “Trump” and “Racist” connection. And all the usual nonsense that you would expect to happen is happening.
Scott Stephenson, Founder and CEO of Deepgram had this comment:
“AI is only as smart as the data it’s trained on. Voice recognition should be about understanding, not assuming. This is a reminder that companies need to constantly refine their models to avoid bizarre and potentially harmful mix-ups. The goal isn’t just accuracy, it’s trust.”
“Voice AI is exploding because, let’s face it, talking is faster than typing. But speed means nothing without precision. If AI can’t truly understand what we’re saying – accents, slang, emotions – then it’s just noise. The next wave of innovation is about making AI listen smarter, not just faster.”
On one hand, I am surprised that this wasn’t caught in QA. But on the other hand, I am not surprised because how far do you go to test a speech to text engine like this one? You can’t test every single word or name out there. Thus you can expect something like this to happen again. And chances are there’s nothing nefarious going on.
Share this:
Like this:
Related
This entry was posted on February 26, 2025 at 12:26 pm and is filed under Commentary with tags Apple. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.