The next evolutionary step in machine translation

Seven decades from the very first trials of machine translation, critics claim that we have yet to reach human parity. New research suggests that “Responsive MT” could evolve to deliver increasingly appropriate results.

Text by Arle Lommel


Image: © Tashatuvango/

Today, machine translation (MT) enables organizations and individuals to access ever-increasing volumes of international content, but it still suffers from major limitations. For MT to move forward, it requires substantial improvement in the intelligence of the data used to train it, as well as in the quality and relevance of its outputs. CSA Research predicts that current developments and trends in Artificial Intelligence will result in what we term “Responsive Machine Translation.” As Responsive MT begins to roll out, production managers at language service providers (LSPs), translators, and reviewers will need to interact and prepare for new business opportunities. On the content creator side, information architects, technology leads, and localization managers will all benefit from understanding how MT can change and improve their deliverables.

In the course of seven decades of ...