AI and Online Tools – Is Your Private Information Really Private?

Very few people would argue with the assertion that AI has revolutionised our daily lives – and mostly for the better. It’s hard to imagine a world without online shopping, smart home appliances, or social media. Virtual assistants remind us of diary appointments and soothe us with our favourite playlists. Apps save us time with shortcuts to frequently visited websites, while spam filters help keep our inboxes secure. From automated passport control to challenging chess games, we’re in constant communication with our digital doppelgangers.

The workplace has similarly been transformed by breakthroughs in robotics, search optimisation, information retrieval and data management. In my own industry, AI has not only helped to streamline the translation process, but has also democratised access to high-quality language services and expertise.

A little knowledge can be a dangerous thing, however – and we’ve seen time and again in the language industry that there are no short cuts to professionalism, reliability and trust. At best, the results will be unsatisfactory; at worst, you could be putting sensitive data at risk.

I was reminded of the importance of trust, in particular, when I read a recent article in Digital Journal warning of the downside of using free online tools.[1] The article was prompted by the news that Microsoft contractors have admitted to listening in on Skype calls with the stated aim of improving their automated translation service, which ‘learns’ through human intervention. The same accusation has been levelled at Google Assistant, Apple’s Siri and Amazon’s Alexa. This might all seem innocent enough, but we need to bear in mind that not all the material being transcribed is harmless chitchat; moreover, the vast majority of users probably don’t realise that their interactions could be monitored (in spite of the usual warnings in the small print).  

As far back as 2012, Dan Gillmor, director of the Knight Center for Digital Media Entrepreneurship at Arizona State University, was raising concerns about privacy issues for Skype users.[2] Yet it seems the message is not getting through to people on either a personal or a professional level. I wonder if the 8,000 callers per year to the Skype NHS team would be so quick to share their medical problems online if they knew their confidentiality could be in jeopardy.[3]

It is vital, therefore, that businesses who use AI-enabled translation tools fully understand the potential compromises involved for the privacy and security of their own interactions, not to mention those of their clients. I remember a case a couple of years ago when one of our customers was obliged to formally block Google Translate after an employee, without thinking twice, ran a sensitive client document through the tool. Not only is this a breach of client confidentiality, it is also now in contravention of GDPR – and in a worst-case scenario it could lead to serious cyber fraud. The best way, therefore, to circumvent the risk of unwittingly exposing sensitive or identifying data is to ensure that any translation tool your business uses is GDPR-compliant as standard.

At Guildhawk we’ve always believed there’s no substitute for the personal touch, and the vast majority of our work is still undertaken solely by human linguists. There are occasions, however, when AI-assisted translation can offer a quick, convenient, cost-effective alternative to our human translation service. We knew some of our clients would benefit from such a resource. But when we looked around for a suitable GDPR-compliant tool, we drew a blank – so, typically, we set about customising one ourselves! GUILDHAWK AIDED is the result. Like all our services, it’s fully backed by our ISO27001 certification for data security. And we still add that all-important personal touch by building in detailed human reviews and quality checks – our guarantee that the finished product meets the same scrupulous standards as our regular translation service. This allows our clients to tick all the boxes – they can embrace the changes and benefits that AI can provide, while being confident that their data (and their client’s) is completely secure.

AI is here to stay and, in the words of our CEO Jurga, it is not something to feared.  We at Guildhawk are excited about where technology is going to take us next in our mission to help increase international trade and productivity, and are committed to ensuring we apply our non-negotiable ethics and security standards to anything we take on in this exciting arena.

1. http://www.digitaljournal.com/tech-and-science/technology/be-careful-what-you-say-microsoft-listens-in-on-skype-calls/article/555525
2. https://www.theguardian.com/commentisfree/2012/jul/27/skype-encyption-methods-police-access
3. https://www.digitalhealth.net/2019/06/skype-doctors-ae-visits-manchester/