World Affairs- Jan 26, 2024
- 1,774
- 1,019
- Origin
- Residence
- Axis Group
- Copy to clipboard
- Moderator
- #1
Source
:
https://www.theatlantic.com/technology/archive/2024/03/generative-ai-translation-education/677883/
Thanks to AI, people may no longer feel the need to learn a second language.
No machine can replace such a profound human experience. However, tech companies are turning automated translation into more products. As technology normalizes, we may find that deep human connections have been allowed to be replaced by ones that are technically skilled but ultimately hollow. AI-based language tools are now available in social media apps, messaging platforms, and serial streaming sites. Spotify is experimenting with voice generation tools from chat GPT maker Open AI to translate podcasts with its own voice while Samsung claims that its brand-new Galaxy S24 smartphone can translate phone calls simultaneously. Meanwhile, online gaming platform Roblox claimed last month that its AI-based translation tool is so fast and accurate that English-speaking users of that platform may not realize that their conversation partner, for example, is actually South Korea.
The neural networks of machine learning systems that provide intelligence-driven programs such as Chat GPT have improved the quality of automated translation rapidly over the past few years, making even older tools like Google Translate much more accurate. At the same time، the number of students studying foreign languages in the United States and other countries is declining. According to the latest data from the Modern Language Society, more commonly known as MLA, total enrollment in non-English language courses at American colleges decreased by 29.3 percent from 2009 to 2021. In Australia، only 8.6% of high school students studied a foreign language in 2021، one of the lowest historical levels in the field.
The technology, which works particularly well for high-resource languages such as English and Chinese, and less for languages like Swahili and Urdu, is also used in much more risky situations such as translating refugee statements and first-hand accounts from conflict zones. Musicians are now using it to translate songs."
"One of the most important things is using a startup called Jumpspeak, which builds a language learning app similar to Duolingo and Babbel. Instead of hiring real bilingual voicemails, Jumpspeak appears to have used people generated by artificial intelligence to read translated scripts in at least four ads on Instagram and Facebook. At least some of the characters shown in the ads appear to be default characters on the Hey Gen platform. "I've been trying to learn languages all my life. Then I learned Spanish within six months, got a job in France and learned French. I learned Mandarin before visiting China." These words are said by an artificial avatar in one of the ads as it switches between all three languages. Even a language-learning program at least succumbs to the appeal of artificial intelligence in its marketing. Alexander Wojka, a communications specialist who works for another AI video production service, tells me that when he was looking for a program to teach his children Romanian, he was confronted with an ad for Jumpspeak, the language that their grandparents speak.
He argued that the ad showed how deepfakes and automated translation software could be used to mislead or deceive people. The world now sees how this process can go wrong. Earlier this month, a far right conspiracy theorist shared several videos created by artificial intelligence on Twitter of Adolf Hitler's 1939 speech in English instead of the original German.
The videos, apparently produced using software from a company called Eleven Labs, recreate Hitler's voice. It was a strange experience to hear Hitler speak in English, and some people published comments that showed that it was easy to sympathize with him. For example, one user wrote, "This person seems to care more about his country than anything else." Gabriel Nicholas, a researcher at the nonprofit Center for Democracy and Technology, told me that part of the problem with machine translation programs is that they are often viewed as neutral.
Deborah Cohen, a professor of Spanish and Portuguese at Indiana University Bloomington who has worked on the importance of language learning to boost U.S. national security, told me, "Students ask, "How do you say that in Spanish?" And I say, "You don't just say it in Spanish." I recently came across a beautiful and particularly illustrative example of this fact in an article written by a translator in China called "Anne." "Building a ladder between languages such as Chinese and English is sometimes as hard as a surgeon to bridge the heart," he writes.
The metaphor seemed a little strange at first, but luckily I didn't rely on Chat GPT to translate the words from the original Mandarin. I was reading a human translation by a professor named Jeffrey Ding, who was helpful to point out that the analogy might refer to a type of heart surgery that has recently become popular in China. This may seem like a small detail on the surface, but understanding the context of its application brought my analogy much closer to what the phrase was trying to say. However, most learners will never achieve anything close to fluency necessary to determine whether a translation is close enough to the original. If the professors accept that automated technology will far outpace the average Russian or Arabic technical skills, their focus will ideally shift from grammar practice to developing cultural competence or understanding the beliefs and practices of people from different backgrounds.
Jen William, Dean of the School of Languages and Cultures at Purdue University, said instead of cutting language courses in response to artificial intelligence, "Schools need to emphasize more than ever the intercultural components of language learning that are very useful for students attending those classes." The reality is that it may be possible to translate through technology, but it is not possible to understand the metaphors of the culture associated with that text or phrase. "Something like a global translator is now in development, but it lacks the power to bridge cultural divides in a way that humans can."
No machine can replace such a profound human experience. However, tech companies are turning automated translation into more products. As technology normalizes, we may find that deep human connections have been allowed to be replaced by ones that are technically skilled but ultimately hollow. AI-based language tools are now available in social media apps, messaging platforms, and serial streaming sites. Spotify is experimenting with voice generation tools from chat GPT maker Open AI to translate podcasts with its own voice while Samsung claims that its brand-new Galaxy S24 smartphone can translate phone calls simultaneously. Meanwhile, online gaming platform Roblox claimed last month that its AI-based translation tool is so fast and accurate that English-speaking users of that platform may not realize that their conversation partner, for example, is actually South Korea.
The neural networks of machine learning systems that provide intelligence-driven programs such as Chat GPT have improved the quality of automated translation rapidly over the past few years, making even older tools like Google Translate much more accurate. At the same time، the number of students studying foreign languages in the United States and other countries is declining. According to the latest data from the Modern Language Society, more commonly known as MLA, total enrollment in non-English language courses at American colleges decreased by 29.3 percent from 2009 to 2021. In Australia، only 8.6% of high school students studied a foreign language in 2021، one of the lowest historical levels in the field.
The technology, which works particularly well for high-resource languages such as English and Chinese, and less for languages like Swahili and Urdu, is also used in much more risky situations such as translating refugee statements and first-hand accounts from conflict zones. Musicians are now using it to translate songs."
"One of the most important things is using a startup called Jumpspeak, which builds a language learning app similar to Duolingo and Babbel. Instead of hiring real bilingual voicemails, Jumpspeak appears to have used people generated by artificial intelligence to read translated scripts in at least four ads on Instagram and Facebook. At least some of the characters shown in the ads appear to be default characters on the Hey Gen platform. "I've been trying to learn languages all my life. Then I learned Spanish within six months, got a job in France and learned French. I learned Mandarin before visiting China." These words are said by an artificial avatar in one of the ads as it switches between all three languages. Even a language-learning program at least succumbs to the appeal of artificial intelligence in its marketing. Alexander Wojka, a communications specialist who works for another AI video production service, tells me that when he was looking for a program to teach his children Romanian, he was confronted with an ad for Jumpspeak, the language that their grandparents speak.
He argued that the ad showed how deepfakes and automated translation software could be used to mislead or deceive people. The world now sees how this process can go wrong. Earlier this month, a far right conspiracy theorist shared several videos created by artificial intelligence on Twitter of Adolf Hitler's 1939 speech in English instead of the original German.
The videos, apparently produced using software from a company called Eleven Labs, recreate Hitler's voice. It was a strange experience to hear Hitler speak in English, and some people published comments that showed that it was easy to sympathize with him. For example, one user wrote, "This person seems to care more about his country than anything else." Gabriel Nicholas, a researcher at the nonprofit Center for Democracy and Technology, told me that part of the problem with machine translation programs is that they are often viewed as neutral.
Deborah Cohen, a professor of Spanish and Portuguese at Indiana University Bloomington who has worked on the importance of language learning to boost U.S. national security, told me, "Students ask, "How do you say that in Spanish?" And I say, "You don't just say it in Spanish." I recently came across a beautiful and particularly illustrative example of this fact in an article written by a translator in China called "Anne." "Building a ladder between languages such as Chinese and English is sometimes as hard as a surgeon to bridge the heart," he writes.
The metaphor seemed a little strange at first, but luckily I didn't rely on Chat GPT to translate the words from the original Mandarin. I was reading a human translation by a professor named Jeffrey Ding, who was helpful to point out that the analogy might refer to a type of heart surgery that has recently become popular in China. This may seem like a small detail on the surface, but understanding the context of its application brought my analogy much closer to what the phrase was trying to say. However, most learners will never achieve anything close to fluency necessary to determine whether a translation is close enough to the original. If the professors accept that automated technology will far outpace the average Russian or Arabic technical skills, their focus will ideally shift from grammar practice to developing cultural competence or understanding the beliefs and practices of people from different backgrounds.
Jen William, Dean of the School of Languages and Cultures at Purdue University, said instead of cutting language courses in response to artificial intelligence, "Schools need to emphasize more than ever the intercultural components of language learning that are very useful for students attending those classes." The reality is that it may be possible to translate through technology, but it is not possible to understand the metaphors of the culture associated with that text or phrase. "Something like a global translator is now in development, but it lacks the power to bridge cultural divides in a way that humans can."
