Automation hits multilingual workforce with real-time translation software for businesses

With the rise of chatbots and automated phone systems, telephone-based customer service roles have seen increasing automation. One skill that has protected many workers from having their jobs replaced by software, however, is the ability to speak multiple languages.

But while software-based translation has traditionally paled in comparison to a multilingual human being, that’s beginning to change. The first gadgets allowing humans to carry out a conversation across two languages are starting to be launched, and even Google’s own free services are getting in on the act.

Now, however, companies are likely to follow suit, with the launch of a software suite for companies in need of multilingual support for helpdesks and service support.

Developed by Lionbridge Technologies, the snappily named GeoFluent for Enterprise Service Management is a package of business software that includes over-the-phone interpretation for over 350 languages, as well as a virtual translator for digital communications such as email and chat and a self-service document translator.

As a result, it effectively eliminates the need for major companies to hire multilingual workers for any form of customer service, instead replacing their skills with software.

Customer service has seen significant automation in recent years, but multilingual workers have until now been relatively safe

For businesses, the software’s main benefit is its ability to save them money, by not hiring staff to speak multiple languages or offer translation services.

“Service desk agents have historically had limited options to deliver multilingual support. Customised real-time translation technology is an increasingly important piece of IT service management solutions, providing a high-quality, cost-effective alternative to hiring bilingual staff,”  Robert Young, research director of IT service management and client virtualisation software at International Data Corporation, said in a press release about the software.

“It’s a significant challenge for service desks to communicate effectively across languages, channels, geographies, and time zones. GeoFluent for Enterprise Service Management eliminates that complexity by leveraging unified interpretation and custom-trained AI-based translation to existing communications platforms and channels,” added Tom Tseki, Lionbridge’s vice president and general management, customer care solutions.

“This platform and channel-agnostic approach allows service desks to cost-effectively eliminate language barriers wherever they exist.”

Manufacturing was one of the first industries to see significant automation in the modern era, but other roles are increasingly being affected

While those watching the technology’s developed will unlikely be surprised by the development, for many it will come as a surprise that multilingual skills can be so completely automated.

However, it is yet another example of skillset that not long ago was thought to be completely safe from the oncoming march of automation, but which is now under threat.

Many journalists, for example, are feeling less confident about their own roles with the advent of news-writing bots, while AI composers are likely to be raising some concern in the creative industries.

198 million Americans hit by voter records leak should get immediate credit freeze: experts

The 198 million US voters whose personal data was left on an unsecured server for anyone to access should request an immediate credit freeze to avoid having their identities stolen as a result of the breach, security experts have said.

“The members of the electorate involved in this incident should immediately request a credit freeze with the major credit bureaus, and keep close track of account activity through commercial credit monitoring services, or monitoring of their own accounts,” advised Robert Capps, VP of business development at NuData Security.

The data, which includes personal data and information on who each person is set to vote for and why, is thought to be the largest ever exposure of voter data, covering the vast majority of the 200 million people registered to vote in the US.

It was left on an open Amazon S3 storage server by Deep Root Analytics, a Republican data analytics company, and was discovered by Chris Vickery, a cyber risk analyst from UpGuard.

At present there does not appear to be a way in which individuals can check if they were affected, but anyone registered to vote in the US is likely to be at risk.

Graphic courtesy of UpGuard

While the focus of the data was voting behaviour, containing information on the subject that goes back over a decade, voters should be more concerned about how their data could be used for more malicious purposes.

“This is a serious data leak, which allows nation states to target ordinary US citizens for additional attacks and surveillance, as well as detailed voting information,” said Capps.

“If this wasn’t bad enough, this highly detailed data could potentially be combined with stolen personal data from other data breaches already available on the dark web to create rich profiles of these individuals.

“Such profiles can be leveraged by cybercriminals and nation-state actors to not only track voting habits, but also use their identities for account takeovers, apply for new credit, and much more.”

People cast their votes in the 2012 presidential election in Ventura Country, CA. Image courtesy of Joseph Sohm / Shutterstock.com

While the risk to those affected is similar to previous leaks, this is not a leak or hack in the classic sense, but instead a matter of poor security practices.

“It sounds to me that this is another case of incorrectly secured cloud based systems,” explained Terry Ray, chief product strategist at Imperva.

“Certainly, security of private data – especially my data, as I am a voter – should be of paramount concern to companies who offer to collect such data, but that security concern should ratchet up a few marks when the data storage transitions to the cloud, where poor data repository security may not have the type of secondary data centre controls of an in-house, non-cloud data repository.“