- iOS changes “trump” to “racist” during transcription
- Apple says the bug is now corrected
- Official explanation is “phonetic overlap”
The owners of iPhone have noticed a particular bug in recent days: “Trump” automatically correctly to “racist” when using the text dictation mode. According to Apple, this is a problem with “phonetic overlap”, and a fix is already in preparation.
After the videos of Tiktok from the Slip have become viral, Apple provided a statement to the Guardian and others, blaming the “phonetic overlap” between the two words: “We are aware of a problem with the voice recognition model Who feeds the dictation, and we take a correction as soon as possible, “said a spokesperson.
While many people were able to recreate the Blip, it did not happen every time – and the text seemed to return to “Trump” after a short delay. The latest online reports suggest that Apple’s fix has already taken effect, you may not see it anymore.
In its explanation, Apple suggested that its vocal recognition engines were struggling to distinguish between words with “R” in them. Other tests have suggested that iOS has not always obtained the word “racist” either, although the engines of Apple’s word to the text of Apple were very reliable.
‘Just not plausible’
Apple wishes to trace a line under this as soon as possible and have the error correct. It seems particularly regrettable that a transcription bug like this one would link two specific words to trigger a wave of controversy and politically loaded debate.
Peter Bell, professor of speech technology at the University of Edinburgh, told the BBC that Apple’s explanation was “simply not plausible” given what we know about speech technology. “It probably shows someone who has access to the process,” said Bell.
John Burkey, founder of Wonderrush.ai, gave a similar option to the New York Times: “It feels a serious farce,” he said. “The only question is: Did anyone slip this into the data or slip into the code?”
This also feeds the wider conversation on the AI and its reliability, because the AI models are used to convert the word spoken into a transcribed text – something you can now do on any modern smartphone. Whether it is to meet notes or show subtitles, we must be able to rely on the precision of this technology to spread rapid.
