January 08, 2020
By Blog Team
Ever since Amazon released its smart speaker Alexa all the way back in 2014, some businesses have begun to believe that the role of interpreters may be null in the coming years. What's more, Google’s new version of the Google assistant interpreter function has been ‘perfected’ just in time for the release of the Google Nest home hub.
Can Google Nest compete?
For those of you who are unfamiliar with the product, Google Nest is one of Google’s latest attempts to become the leader of the AI assistant market. It’s the result of merging Google’s Home Hub and Nest products and offers a range of AI services for your home. Google Nest is connected to YouTube, and offers calendar alerts as well as ‘smart home’ capabilities like controlling your lights, plugs and other devices.
But one of the most talked about features is Nest’s interpreter mode. The newest version of the interpreter mode translates spoken languages in real time. The device offers this service for 29 languages, such as Spanish, Mandarin and English, short of the 100 languages on the earlier Google Translate service.
The interpreter mode on devices like Nest was created with a modest initial goal: To enable meaningful face-to-face conversations between speakers of different languages.
With the recent announcements that New York’s JFK Airport and American Airlines lounges will begin trialling Google Nest Hubs as interpreters for tourists, there is growing evidence to suggest that Google is looking beyond the home and eventually may aim to compete with human interpreters in the future.
Michael Bronstein, Vice President of Product for Google Assistant, said:
"It's very futuristic. At the beginning, it's going to be basic things like, 'Hey Google, play music,' and 'Turn on this light.
"But as you see with translation and so forth, I think the complexity of the tasks [that interpreter mode] is going to be able to handle will increase.’’
Google's technology was offered at concierge desks in America as a test run. The results of this trial were interesting. Although the mistranslation from a ‘shrimp allergy’ to a ‘sand allergy’ is portrayed humorously in the video below, a less obvious mistranslation could have fatal consequences for a guest.
What do human interpreters think about AI assistants?
Human interpreters have mixed opinions on the developing technology. Most interpreters feel safe in their position, knowing that businesses will initially be reluctant to pit a potentially faulty service against trusted interpreters.
Others are more worried about the impact of AI assistants on their work. While interpreters are keenly aware that humans are necessary for quality, professional interpretation, they may feel at a loss to compete with AI’s glossy marketing and one off price.
But do we think there’s anything for them to worry about?
If there’s one-word all businesses are wary of, it’s risk. Relying on AI for your interpretation services poses a number of risks – and most of them could have serious negative consequences. For one thing, there’s the usual financial risk that you have to worry about. It’s been suggested that language and communication mistakes cost businesses around $2 million just in the US!
While financial mistakes can easily make or break a business, there are more troubling risks as well. Interpreters are a ubiquitous presence at high-stakes international meetings and conferences. Whether they are sitting in a meeting for a proposed international merger or a diplomatic mission, they are entrusted with some of the most important data and secrets in the world.
In cases like these, it would be unwise to trust AI interpreters. AI assistants are known for several high-profile data breaches since they came onto the market. That’s a risk that simply cannot be taken in such important situations. AI assistants could be the ultimate ‘fly on the wall’ for malicious hackers. That’s why businesses and governments in high stakes situations should always trust a thoroughly vetted, highly experienced interpreter. AI simply can’t compare in this regard.
Even in the everyday world of business, every corporation ought to be thoughtful about data security. The interpretation of conversations involving finances, acquisitions, and HR data could all be weaponised if AI goes wrong. It could even dismantle the business, as average cyber-attack now costs a company around $200,000 – enough of a cash burden to shut some SMEs down for good.
Nuance in AI assistants
To be a good interpreter, you need to understand the power of metaphors and idioms in speech. Interpreters are trained to use prior research, knowledge or context clues to discern what their client is saying. What's more, they are usually fully briefed before each job, allowing them insight into the kind of language they are likely to encounter.
On the other hand, AI assistants are far more literal. They may be blessed with a shiny interface and breezy marketing campaign, but they aren't known for their powers of nuance.
For example, in a meeting, an Italian businessperson may wish their English-speaking business partners would stop skirting around an issue in a stereotypically-British fashion and ‘‘non avere peli sulla lingua.’’ An interpreter would understand that this idiom is encouraging the English-speaking partners to be more straightforward and get to the point.
If that same businessperson decided to instead use an AI assistant like Google Nest, there’s a high likelihood that the AI would incorrectly interpret the idiom. In the end, the English-speaking guests could all be left scratching their heads, and perhaps slightly offended, wondering why the Italian had said something about hairs on their tounge. (The literal translation of the phrase ‘‘non avere peli sulla lingua" is "not to have hairs on the tongue," with the English equivalent being "Not to mince one's words.")
It isn’t hard to think of situations where this machine interpretation could be very awkward. AI can also prove disastrous when translating homonyms or words with multiple meanings. For example, the word ‘set’ has 464 meanings. Can you really trust AI’s snap judgement of which one would fit the situation best?
Interpretation isn’t just about changing words from one language to another. At its core, all interpretation is about communication.
Fundamental to communication, particularly in business, is nonverbal language, interpretation and interpersonal qualities. AI assistants simply cannot offer any of these attributes. Interpreters provide the ‘human’ element that can persuade, create empathy and promote understanding between all parties even in the tensest of situations.
From our point of view, that's one major reason that AI assistants will never replace human interpreters.
Complex projects require complex solutions.
AI translators and devices are in their early stages. While they may be a worthwhile investment in low risk situations when faced with simple communication scenarios, more complex projects are currently out of their grasp.
You’ll need to research the options available and find out what’s right for your business.
Wolfestone: Your interpreting partner
Are you looking for a conference interpreting partner who can support you during your business meeting abroad? Or would you like to invite your foreign business partners to your UK offices? Wolfestone is here to help take the stress out of your multilingual meetings, appointments and conferences.
Alongside simultaneous and consecutive interpreting, we also offer a telephone interpreting service– a quick, easy and affordable way to meet your immediate interpreting needs.
Why not get in touch and discover how we can help you?