Our Analysts' Insights

Artificial intelligence has become the new buzzword in the language services industry, with countless technology vendors and LSPs proclaiming that their AI efforts will revolutionize the field. Providers are scrambling to keep up with the fast pace of development to understand how it will affect their business and how to remain relevant in a world full of automation. Buyers are trying to figure out what it means for both their own processes and when they purchase services from vendors. Humans at all levels of the supply chain question what AI will do to their jobs.

And as you might expect, amid all this buzz, there is a lot of confusion. The majority of it ties to what exactly is artificial intelligence. AI in the form of machine learning and predictive analytics represents an evolved form of automation, yet people across the industry indiscriminately use the word to refer to “automation” in general. For example, rule-based machine translation software or a preconfigured workflow in a translation management system are definitely automation – but they are not AI. Instead the rules in these systems were meticulously created by humans based on what they considered best practices.

For automation to qualify as AI, it requires deep learning whereby computers build a mathematical model from sample data, analyze it for statistical patterns, and make predictions or decisions based on the training data. With AI, it’s machines that create rules to apply to automate work. Over time, they learn from outcomes to deliver better results. Examples of AI include neural machine translation and self-learning systems that identify problems in translation memory segments.

AI affects organizations on different fronts: language with neural machine translation and various tools to clean up language resources, production with expert systems that handle project and vendor management functions, business development with marketing, sales and customer service tools, and business operations through advanced automation in areas such as human resources.

However, in the end, whether new advances are rule-based systems developed by humans or machine learning by sophisticated systems, the reality is that automation is permeating every aspect of language services. Why? Besides the fact that the AI form of automation is au courant, it’s also essential. LSPs need to remove as much waste, inefficiency, time, and cost from their operations – and automation of any type is a major step in meeting those goals. For example, just a bit more than half of LSPs we surveyed have adopted machine translation in their production model. Everyone has a vested interest in learning about developments in the field so they can adapt and react – with their processes, their business, or even their careers.

Join our presentation on Debunking AI Myths at GALA in Munich, Germany next month. Our session will help you reframe the dialog around the right questions to ask without getting tangled in misleading messages you hear in press releases, the twittersphere, and in the business press.