Artificial intelligence isn’t science-fiction anymore. This rapidly-developing technology now has numerous applications in a wide range of industries. The translation industry, in particular, is benefiting from the rise of AI. To understand how it first helps to review the technology’s basic history.
The 1950s - Innovations like the Ferranti Mark 1’s chess program introduced the world to the concept of “narrow AI.” This term refers to basic AI that can follow rules like a human.
The 1990s - During the 1990s, machine learning began playing a more substantial role in the development of AI technology. Through machine learning, algorithms teach computers to separate data into relevant subsets. The AI can theoretically review the subsets of data and make decisions accordingly.
The 2000s - The early 2000s saw the rise of deep learning, a branch of machine learning that employs algorithms designed to mirror the human brain. AI developed with deep learning should, like a person, be able to solve new problems by learning from past experiences.
The Future - It won’t be long before general AI is much more commonplace. General AI refers to machines that are able to perform a wide range of intelligence-based tasks the same way a human would. Think of an autonomous robot from your favorite sci-fi film, and you get a basic sense of what general AI is.
The Role of Deep Learning in the Translation Industry
For those working in the translation industry, deep learning (and its applications in the development of AI) has been revolutionary. Anyone who has ever used an online translator only to receive a finished translation with awkward sentences and numerous errors knows that narrow AI is very limited in this field.
Narrow AI translators can follow a basic process of separating a sentence into fragments and looking up foreign words in a dictionary, but it can’t pick up on the nuance of language. If you’re attempting to use this technology for learning or marketing translation, for example, this can greatly impact how effective your product is and possibly damage your reputation if the quality of the translation suffers. The same consequences apply across all industries.
A deep learning AI translator is more likely to correctly identify sarcasm, get a joke, or even appreciate the cultural and situational context. At Google, innovators have recently demonstrated just how effective this technology can be.
Google’s Breakthrough
Up until recently, Google Translate relied on narrow AI. In 2016, however, the tech giant experimented with a new approach to online translations. Using a new system called Google Neural Machine Translation, researchers explored the limits of what AI translators can do.
For instance, could an AI accurately translate a new language pair without additional training? In other words, could an AI that was trained to translate Japanese into English and Korean into English also translate Japanese into Korean with no extra programming?
Apparently, yes. While narrow AI, which essentially performs word-to-word translations, would be unable to achieve this feat, Google’s new AI taught itself to translate Japanese into Korean by creating its own artificial language consisting of common elements in both languages.
Applications
It’s easy to imagine how this technology might be applied to a wide range of industries. For example, law firms frequently involved in international litigation can speed up the process of resolving a case by easily translating documents for review. A professional human translator should definitely confirm the accuracy of any documents that are to be signed or delivered, but AI can still help when a legal team needs to sort through reams of documents in another language to find relevant details.
Any business translating its website or marketing materials will also benefit. True, as with law firms, businesses must consult with translation professionals to ensure all content is localized for the culture of a given audience. That doesn’t mean they can’t rely on AI to provide the basic framework for translated materials.
Essentially, while AI won’t replace people in the translation industry, it will allow translation firms to work more efficiently for their clients. Professionals in any industry that requires translations should pay attention to these developments.
x
