Artificial Intelligence: a key driver of language services

Artificial Intelligence: a key driver of language services

Artificial intelligence (AI) is commonly described as intelligence demonstrated by machines, as opposed to the natural intelligence displayed by animals and humans.

In the context of human language, it is closely linked to Natural Language Processing (NLP), which allows machines to read and understand what is being read. One of the straightforward applications of this concept is machine translation.

Traductanet attends LocTalk session on emerging trends in AI

On 27 October, Traductanet joined the industry’s biggest online event on translation and localisation, LocTalk by Smartcat, and learned more about how AI and machine translation are not new and how we are only scratching at the surface of what is possible.

Speakers discussed the latest innovations, how companies can make more use of the technology, how to keep up to date with the latest advances, and how to balance the human v machine component.

Keep reading to learn more about AI’s uses in machine translation, text-to-speech solutions and even interpreting, as well as other applications.

AI beyond mere machine translation

According to Josef Kubovsky, CEO of Nimdzi, an international market research and consulting firm advising on language services, localisation programs, business practices and mergers and acquisitions, the main opportunity for AI is that it is helping us with the content and is found throughout the process.

“Speed, in my opinion, is a major factor to this [task automation], and we are definitely heading towards an amazing future that will help us go farther […] People are getting used to receiving their content in their native language and it is natural for them.”

Research on the topic certainly seems to confirm this statement, with 72.1% of internet users saying they prefer to visit websites in their native language and 78% of online shoppers reporting they are more likely to purchase if a website is localised.

Text-to-speech, lip dubbing and interpreting

Derick Fajardo, who works on international strategy and machine translation at Harvard Business Publishing, agrees. He explains that the simplest application for AI is machine translation, adding that it can be used for so much more: speech automation, lip dubbing, simultaneous interpretation, quality assurance and pricing, etc.

The panellists spoke about how it will soon be possible to videoconference with someone speaking English, for example, while their voice is automatically dubbed into Japanese in real time. This will enhance global communication and enable greater understanding between parties.

Researchers in India have even developed a new Generative Adversarial Network (GAN) called LipGAN, which can match the lip movements of translated text in an original video and correct them in the dubbed version.

Coupled with this tool, speech-to-speech translation forms the whole model with the following pipeline: the model uses speech recognition to transcribe the speech in a video; it then uses a trained specialised model to translate the text; finally, the speech recognition model converts this text into the spoken word.

Josef Kubovsky also says the industry has a duty towards linguists in terms of fair pay and suggests paying the voice owners every time their voices are used in a dubbed video.

Next steps in the evolution

Harnessing AI to achieve other business goals is also a growing trend within multilingual companies, which comes down to volume.

Budgets and time often constrain companies, so translating everything in good time is often impractical. Being able to reach many regions and people with the simple push of a button can be very rewarding; however, much remains to be done regarding the usability of tools.

As for other areas, AI can also be useful for project management and recruitment, testing qualified suppliers and profile sorting. This is something with which Traductanet is very familiar.

We use several systems and resort to specific partners to manage the entire translation process automatically and efficiently, freeing our project managers to focus on the customers.

Do not hesitate to contact us and request a quote if you require a language service provider that will help drive down costs and increase delivery times while offering high-quality texts.