Back in the late 1900s, Bill Gates thought he could accumulate people from different areas of the world on the digital platform. These people, speaking 7,000 different languages can gather on the internet like they would at a town square.
And he was right! The World Wide Web has definitely made it possible for people to interact with each other without any physical barriers.
However, a recent study indicates a challenge to this blessing.
According to a study by Amazon Web Services and the University of California, the majority of the translations present in the Internet database are not up to the mark. The study compared over 6 billion sentences translated in at least two languages to assess the quality. They concluded that the more the sentences were translated, the worse they got.
Researchers predict that the low-quality translation was most probably conducted by computers. They also said that these computers are writing new stuff, especially when it comes to languages that are not as widespread. For example, Woolf and Xhosa from Africa.
Unfortunately, the lack of resources on the internet has led many users to rely on these bad translations. Additionally, many of these translations have led to funny and embarrassing scenarios for the users.
For example, Google once translated "Russia is a great country" into something about a fictional place in "The Lord of the Rings." In 2019, Facebook's translation tool also made a big mistake with the Chinese President's name in an article translated from Burmese. After realizing the mistake, Facebook did issue an apology, citing technical issues for the error.
And there was a funny mistake when translating medical advice for Armenian speakers. Instead of suggesting "ibuprofen for pain," it said to take an "anti-tank missile for pain."
Seems like most of the language translation services are simply pushing content on the internet just to make money from ads. What do you think?
Photo: AI-gen
Read next: Meta's Facebook and Instagram Are the Most Data Hungry Apps According to This Study
And he was right! The World Wide Web has definitely made it possible for people to interact with each other without any physical barriers.
However, a recent study indicates a challenge to this blessing.
According to a study by Amazon Web Services and the University of California, the majority of the translations present in the Internet database are not up to the mark. The study compared over 6 billion sentences translated in at least two languages to assess the quality. They concluded that the more the sentences were translated, the worse they got.
Researchers predict that the low-quality translation was most probably conducted by computers. They also said that these computers are writing new stuff, especially when it comes to languages that are not as widespread. For example, Woolf and Xhosa from Africa.
Unfortunately, the lack of resources on the internet has led many users to rely on these bad translations. Additionally, many of these translations have led to funny and embarrassing scenarios for the users.
For example, Google once translated "Russia is a great country" into something about a fictional place in "The Lord of the Rings." In 2019, Facebook's translation tool also made a big mistake with the Chinese President's name in an article translated from Burmese. After realizing the mistake, Facebook did issue an apology, citing technical issues for the error.
And there was a funny mistake when translating medical advice for Armenian speakers. Instead of suggesting "ibuprofen for pain," it said to take an "anti-tank missile for pain."
Seems like most of the language translation services are simply pushing content on the internet just to make money from ads. What do you think?
Read next: Meta's Facebook and Instagram Are the Most Data Hungry Apps According to This Study
