Fuhgeddaboudit! AI has a hard time with NYC, NJ, LI accents

Hey ChatGPT – are you talking to me?

If Siri and Alexa tend to misunderstand many of your requests, it could be the accent.

A new survey suggests that the distinct intonations of people from New York City, New Jersey and Long Island are among the most difficult for artificial intelligence technology to decipher.

To absolutely no one’s surprise, Southern pulls in at the top of the list as the most difficult accents for AI voice recognition programs to understand, according to language resource site Guide2Fluency.

“Words like ‘y’all’ and ‘fixin’ to’ can be particularly confusing to AI, leading to frequent misinterpretations,” the site said.

New York City’s accent — perfect for buying some gabagool in Brooklyn, but not so great when HE’s looking for capicola — followed in second place, while New Jersey came in third.

The Texas accent ranks fourth hardest, the survey of 3,000 shows.

Shockingly, the Bostonians were ranked No. 5, not first.

“Accents from major urban centers” such as New York City, New Jersey and Boston “often feature distinctive pronunciation and local slang that can hinder AI,” the site explains. “For example, New Yorkers’ tendency to drop the ‘r’ sound (‘cah’ for ‘car’) or ‘pahk the cah’ in Bostonians’ famous Hahvahd Yahd’ can confuse voice recognition systems.

Lawn Guylanders was ranked No. 7. “The Long Island accent, with its nasal tone and long vowels, often leaves AI guessing,” the site explains.

The top 10 rounds out the Appalachian accent (“That feller come a-runnin’ towards ma winder, askin’ for taters”); those from California’s San Fernando Valley (Think Jeff Spicoli’s “All I Need Is Some Tasty Waves, A Cool Buzz and I’m Good”); Miami, an accent influenced by immigrant Cubans; and the people who come from N’Awlins.

Not even the high-ranking residents are spared. The Hudson Valley accent – which is also known as “Albany English” and contains traces of Dutch combined with short New York City vowels – was identified as the 22nd hardest to understand by AI.

Ultimately, the site claims, “these misconceptions have real-world implications,” with users speaking with these accents perceiving AI as “less reliable for everyday tasks, leading to frustration and a lack of trust in the technology.” .