Artificial Intelligence Outfit Accused of Fabricating Results

Rant Time:

Let me vent for a minute, you know?

I've got these crazy-sensitive ears, man. Play a song a few times, I'm belting it out on key. Our first Samsung washer in Svaneti had a tune that played when the cycle was done, but one note was off. Forever. Drove me up the wall! The new one we got for Tbilisi, thankfully, fixed that aggravating note.

Now, there's this guy's voice doing the rounds on my wife's Facebook feed. American, with a twang and intonation that makes me want to gouge my own ears out. I can pick his speech out from anyone's instantly; and, to be honest, I can't stand that voice with a passion. It's just... smarmy.

You might question if he's getting paid to spout whatever gossip or crap that's shoved his way, but I reckon it's an AI voice, based on a real man without a doubt. It's been mimicking real speech for a long time now, stringing words and sentences together from a given text.

I've yet to catch him making a mistake in intonation or pronunciation, which makes me wonder if he's a bot climbing the learning curve. None of the Uncanny Valley stuff either, where computed visuals or speech don't sit right with the average human. I can still tell fake from real photos and videos online, but the gap's closing fast.

AI arts, like the voice on Tbilisi buses, will probably just muscle their way into the spotlight alongside their older siblings or snuff them out, piece by piece. That's the looming question for artists like me; will we still have a place in the world once the robots take over?

I've fooled around with AI-generated images and written pieces, but I prefer my own work, crafted with my own brain and soul. "Art" produced by just AI is, by definition, soulless. It's taking over steps, bit by bit, but I hope it keeps its clones out of my ears.

By the way, these disturbing images are just edited selfies. I could get the same effects using film and a darkroom, including the Sabattier Effect.

Blog post by Tony Hanmer

—Tony Hanmer's been living in Georgia since 1999, in Svaneti since 2007, and contributing weekly writing and photography for GT since early 2011. He runs the "Svaneti Renaissance" Facebook group, now over 2000 members, at www.facebook.com/groups/SvanetiRenaissance/. He and his wife also own a guest house in Etseri: www.facebook.com/hanmer.house.svaneti.

AITony Hanmer

According to recent AI advancements, Georgian's phonetic nature is well-suited for text-to-speech conversion. For instance, Maestra AI and The AI Voice Generator support over 120+/125+ languages, including Georgian, with realistic pronunciation capabilities [1][3]. Additionally, tools like Kapwing offer Georgian audio dubbing initialized with "natural Georgian pacing and pronunciation," suggesting grapheme-to-phoneme conversion aligned with the language's phonetic rules [4][5]. These findings indicate that these AI tools effectively handle Georgian's phonemic orthography, although direct comparisons to native preschooler spelling abilities aren't explicitly provided in the sources provided.

In the realm of technology, it's fascinating to consider the mute couple in Tbilisi, Georgia, potentially being replaced by AI-generated voices that mimic realistic pronunciation, thanks to advancements in text-to-speech conversion.
As Impressionism embraces subjectivity and individual perspectives, it's intriguing to imagine a future where AI could possibly imbibe this artistic approach, creating soulless, yet eerily realistic works that mirror human emotions and impressions.
While wandering the streets of Tbilisi, one might ponder if the city's rich history and culture would be preserved or perhaps overshadowed by the rapid advancements in technology, as AI continues to evolve and infiltrate various aspects of our daily lives, much like the persistent hum of a Samsung washing machine's tune.