Blog

Machine translation: quality isn’t your only concern with using translation technology

Published on October 25th, 2017

machine translation

Normally, when we talk about machine translation, the topic tends to revolve around quality and accuracy compared to professional human translators. This is a constant area of debate – and one that changes as machine translation continues to progress.

However, it’s important we don’t overlook some of the other potential issues with machine translation and similar technologies. Like the recurring problem of security that keeps cropping up with certain automatic translation tools.

 

The security risk with machine translation

While there are always quality concerns with automatic translation tools, the bigger issue businesses should be aware of is security. Everything you type into Google Translate runs through the tech giant’s gigantic data mill. Google will say this is to help improve the quality of its machine translation learning technology, which learns on the go – and that could well be all it gets used for.

Either way, Google Translate is no place for confidential business information or private details that should be kept secure.

The same thing goes for pretty much any online translation tool that lets you type away for instant translations. The security risks are real. In September for instance, The Norwegian Society of Graduate Technical and Scientific Professionals revealed that translations from popular online tool Translate.com were showing up in Google Search.

Notices of dismissal, plans for workforce reductions, contracts and passwords were among the pieces of info available to any Google user around the world. The privacy of individuals and sensitive information from some of the largest companies in Norway were compromised and stock values were even affected as a result of the data leak.

With many machine translation tools, there is a constant security risk. In most cases, it should be a fairly small risk but it’s not one you want to take with sensitive information.

 

Keeping machine translation secure

Security risks aren’t unique to machine translation by any means. The rise of online cloud-based technology brings with it a new world of security risks and its up to software engineers to minimise this in every way they can. However, standards are essentially unregulated and there’s little in the way of legal obligation beyond having a privacy policy, regardless of how secure it may be.

Which means there are machine translation tools that can guarantee security.

You need to do your homework before using any online tool where sensitive information is exchanged. Always start by assuming everything you type will be made public. When it comes to providing information that needs to be secure, guarantee it will be before you use any machine translation tool – especially free online tools.

If you’re taking the agency route, the good news is they are legally obliged to ensure your information remains secure. Talk to your agency about security before you start working with them and make sure there’s something in the contract about them being responsible for ensuring security. Any agency with a decent reputation will have access to the best machine translation tools in the industry and many even build their own tools on a secure network.

Have the conversation before you get started and you should be good to go.

 

Machine translation is a powerful tool when used correctly. Knowing how and when to use it means its shortcomings in terms of accuracy don’t need to hurt the quality of your translation projects. However, it’s also important to understand the security implications many of these tools – and similar online platforms – come with.

In a business environment dominated by big data, leaks are an expensive mistake many can’t afford to make (quite literally). Protecting sensitive information in the digital age is a priority for every business and a few seemingly innocent translations could be all it takes to suffer a leak.

(Visited 200 times, 1 visits today)
Posted on: October 25th, 2017