Our Analysts' Insights

It is incredible to think that, less than eight years after the first publicly available neural machine translation (NMT) systems appeared on the scene, some media coverage already sees NMT as so 2015. As generative AI (GenAI) really exploded into public view in 2022, it wasn’t surprising that an overactive tech press’ imagination would see it as the be-all and end-all of technology. Our recent survey with freelance linguists certainly reflects this view, with many language workers expressing ...

Multimedia, transcribed audio, and AI-generated content in all the world’s digital languages join more traditional content types in filling up data centers. Together they create challenges and opportunities across organizations, raise the alarm for more oversight of content, and further the case for aligning enterprise content strategies, investment, and operations.

Large language models have been in the news a lot in November and December and the coverage has been mixed, to put it mildly. Meta posted its Galactica model on November 15 but took it down just three days later in the face of intense criticism. By contrast, when OpenAI released ChatGPT two weeks later, on November 30, the response was much more positive. Examining why the reactions were so different provides insight into the potential and limitations of machine translation (MT) as well as cauti...

As part of an ongoing investigation into multimedia localization tools and practices, CSA Research is examining enterprises’ global use of video. A combination of professional interest while researching marketing content and personal interest because I’ve just moved, led me to view several TV ads and online videos by international energy providers, including EDF and E.ON. These marketing videos took me down the proverbial rabbit hole, trying to figure out the source and target languages. Which...

Technology developments tend to follow a typical pattern of improvement over time, known as an S-curve. Although it is a familiar pattern, it is worth unpacking its five phases and considering how they apply to language technology and forecasts about it. Examining how they have played out with successive generations of machine translation points to a future in which other advanced natural language processing technologies have tremendous potential to deliver useful and innovative capabilities.

Last week, the Washington Post published an article about Blake Lemoine’s claim that his employer Google’s LaMDA language model/chatbot system had achieved sentience and had a “soul.” Lemoine, an engineer in the company’s responsible AI group, based his assertion on a dialogue in which LaMDA expressed human-seeming sentiments and concepts. Google placed Lemoine on leave, thereby sparking renewed discussion about what machine sentience is and what it means. What can the experience of the lan...

In recent years, there’s been a lot of buzz around “headless” systems – whether for content creation and management or for the translation workflows that feed the global customer experience. The concept being that rather than having a traditional front- and back-end (publishing and creation), these systems allow content to be magically managed, extracted, repurposed, and delivered through a myriad of end points, from mobile apps to corporate websites integrated with a partner’s own custom p...

In October 1991, Unicode 1.0 was first released. In the 30 years since that publication an entire generation of language workers have been educated and started work, never having had to know the “joys” of trying to ungarble text that had gone through multiple encodings. The introduction of Unicode has simplified life for many of us and allowed millions and millions of people to access digital resources in their own languages.

Early in my career, my wife and I lived a few houses down from a truly extraordinary woman named Lorna Call Alder. Born in 1906, Lorna was the first person to develop a now-ubiquitous approach to foreign-language education that emphasizes the deliberate and careful introduction of vocabulary in context and repetition in use within instructional materials. Almost every translator today who has learned a language through formal education owes a small debt of gratitude to this humble woman, who pas...

Our current research into continuous localization shows that the lines have begun to blur between what it means for localization teams to support a traditional Agile model versus one that is more continuous. However, the 53 people who granted in-depth interviews generally agree that the end result remains similar. Organizations must automate as much of the localization process as they can in order to deliver services for mushrooming volumes of content and code that come to them in ever smaller p...