Remember the days of Babel Fish, the translating tool that was all the rage back in the late nineties?
It was the worldwide web’s go-to translation resource. You had to type in a word, and you would get the translated word in the language of your choice. Cut to now, where Artificial Intelligence (AI) has developed to the degree where streaming media can be translated into multiple languages, live.
The internet is truly a thing of a marvel when you consider the cosmic strides it has taken in the last two decades.
Now, in the Covid era, video-calling and virtual event platforms are on top of everyone’s mind. And AI features prominently in this development.
AI has made significant progress in the last few years, irrespective of Covid, to make life a lot more convenient.
Think Siri, Alexa, Cortana, or think song recognizing tools like Shazam. And if you’re a grad student writing an academic paper, consider the lawsuit you can avoid with plagiarism checking tools.
These tools scour billions of pages on the internet to tell you, within a few seconds, if any sentence in your article matches another somewhere in the ether. But like with most advancements, AI has some drawbacks in the world of translation. Let me break it down for you.
How does Artificial Intelligence (AI) play into this? (AI and Language = Machine Translation)
Let’s understand the fundamental difference between translation and interpretation first.
Interpretation is spoken (or conveyed using sign language), and translation is written. Machine Translation (MT) uses AI where a computer takes written text (captions in the case of an online meeting) and automatically translates it into another language with the help of a specially developed algorithm for this purpose.
In a virtual event or conference, machine translation is usually offered bundled with captions for several purposes, including accessibility (for the hard of hearing) and inclusivity (for global audiences).
Other technologies that come into play, in case you want the translation to be available to listen rather than to read, Text to Speech (TTS) is usually added at the end of translation.
Speech Recognition (ASR) is another technology that parses and converts speech to text.
All of these processes – ASR, TTS, MT – rely on AI to run. But like with most other things, none of these technologies is perfect. But technology, as always, will continue to improve and evolve. Since computers essentially process or parse data fed into them, using the algorithm that defines the rules of the process, MT is still a WIP technology.
The translation accuracy is hugely dependent on the quality of the captions in the case of online meetings. But this is also an advantage, as captioning services use human captioners for this purpose. ASR falls short in this area, but human captioning is up to 99% accurate. This linguistic integrity goes a long way in speaking to the AI that powers machine translation.
Are there any shortcomings?
When it comes to online meetings or events, AI falls somewhere in between approximate and accurate delivery.
As I mentioned before, this depends on the quality of the captions rendered through a widget. If ASR, chances of bloopers in the translation are plenty. If captioning is powered by humans, the needle tilts more accurately than approximate. But be warned: this is still not an entirely perfect technology.
The rules of language that MT can adhere to are mostly grammar rules, syntax rules, acceptable spellings (vast differences between the US and UK English, for example, etc. As you know, human communication is complex.
And what AI cannot parse is the finer aspects like inflection, nuance, tone, etc. Professional interpreters are sensitive to these intangible qualities of speech, if you will. Put simply; MT algorithms process the data that they receive. So, the better the quality of captions, the better the end translation will be.
When your online meeting is on a platform that uses a captioning service, where captions themselves are typed out into the text box, live, the quality of the translation notably improves. All this to say that the progress in Machine Translation in the last decade has been tremendous.
AI for your Online Meeting
The online meeting as a communication tool is here to stay. Covid lockdowns have ensured that this is how the business world will collaborate and continue to collaborate in the future. For one, it has kept the motors of the world running.
Anyone with a computer, laptop, tablet, or smart gadget can remain keyed into the happenings of the world and stay connected with their professional network through the last year and a half. And as a way to reach their audiences, virtual meeting companies such as Zoom and WebEx have integrated language technology into their platforms.
And caption providers quickly realized that real-time translation was a problem to solve in this hyperconnected world.
Many have come up with customized solutions, offering a two-prong approach to aid online meetings: captioning and translation. As professional interactions default more and more to online collaborations, this has become the need of the hour.
- Since captioning is usually done by humans, translations need not. There is no need to hire interpreters for your online meeting as AI can take care of translation for you.
- MT is offered bundled with captioning, so you don’t need to pay extra. Hence, it becomes a cost-effective solution if you need the technology but have budget constraints.
- MT will continue to evolve and improve as AI programmers will continue to finetune the process.
- One tool is enough to translate to multiple languages. Some caption providers offer translation capabilities in over 100 languages.
As much as AI is an advantage to your online meeting, there’s still a necessity for the human element.
Having understood that, captioning services now offer machine translation along with captions for virtual meetings. And most are platform agnostic in that they can be integrated into most, if not all, virtual event platforms with a simple widget.
This technology is advancing fairly rapidly, so chances are AI will play into the online world in even more tangible ways in the future.