Many industries use artificial intelligence (AI) to simulate or imitate human behavior. The translation industry has embraced AI in the form of machine translation (MT) to help translate a higher volume of content than ever before. However, MT has not replaced human translation, and there’s little likelihood of that happening anytime soon.
AI Has the Potential to Revolutionize Some Industries
Factory robots use a form of artificial intelligence to replace people, and they can execute well-defined jobs at a high-performance level. They also perform decision-making tasks. For example, if a sensor on the robot identifies a defective part, the robot can order a replacement part. However, that type of decision-making uses defined criteria incorporated into the software driving the robot.
AI can even create new industries. Personal assistants like Apple’s Siri and Amazon’s Alexa have created demand for voice-activated digital assistants. These devices use machine learning to tailor their responses to the owner’s preferences. They “learn” that when the user orders a pizza, the assistant needs to contact the user’s favorite takeout pizza place.
AI's Place in the Translation Industry
There have been many advances in applying AI to language translation. These advances allow computer aided translation (CAT) software like translation memory (TM) and MT to perform translations more efficiently and at a higher quality level, which is a benefit to translation firms and ultimately their clients. The demand for translation is rising, given our global economy and CAT software can add a huge amount of value to the process.
On the other hand, translation software hasn’t risen to the level of understanding the subtlety of all types of language. Also, if the original content was poorly written, the software may not have a defined way to communicate the intended meaning whatsoever.
This issue is giving rise in some parts of the industry for a new role: post-translation editor. Typically, content writers work hand-in-hand with editors to polish, refine and enhance their writing. A post-translation editor can do the same thing for content processed by machine translation.
Some Material Isn’t Appropriate for Machine Translation
There are areas where machine translation isn’t advisable, including most areas of marketing and legal content. Those two areas require an in-depth understanding of more than language rules, and AI hasn’t reached a level that provides that type of insight.
Marketers use their experience, values, judgment, understanding of culture and advanced language skills to prepare marketing content that will resonate with their audience and effectively promote engagement. However, in today’s global economy, that isn’t enough. They must also communicate just as well in a variety of languages.
There are famous examples of situations where poor judgment leading to overtly literal translations made marketing difficult. For example, the slogan "Pepsi Brings You Back to Life" was introduced in China as "Pepsi Brings You Back from the Grave.” KFC also had problems with its slogan in China, where "finger lickin’ good" was translated as "eat your fingers off."
Legal documents pose a similar challenge. Many legal terms have no translation in other languages. Therefore, an accurate translation requires a human translator with legal expertise to bridge the gap. The translator not only must be fluent in the two languages, but they must also have the expertise to compare two legal systems. Learn more
The Cost of Poor Translation
The costs of poor translation can be troublesome or devastating. Consider the costs of these types of poorly translated documents:
- A product user guide: If consumers can’t use the product, the company will see a wave of returns, angry comments on social media and a decrease in sales.
- A medical record: This could result in wasting time to clarify the record or could even cause a misdiagnosis or the wrong medications prescribed.
- A marketing piece: Costs include harm to the company’s reputation, low sales and large investments to correct the error.
- A legal contract: Depending on the error, the cost could be unenforceable contracts, lawsuits, and denial of rights, to name just a few.
How will AI affect translation? Machine translation has its place, but the likelihood that this form of AI will eliminate the need for human translation looks extremely small in the foreseeable future.