How Welocalize’s Darin Goble Sees AI and Large Language Models Reshape Innovation

How Welocalize’s Darin Goble Sees AI and Large Language Models Reshape Innovation

Darin Goble, Head of Solutions & Solutions Engineering at Welocalize, spoke about technology and innovation during SlatorCon Zurich 2023, providing an overview of several aspects of innovation in the context of AI and large language models (LLMs). He also touched on the many unanswered questions about the impact of these disruptive technologies on the language industry.

Goble’s presentation, titled “Digital Transformation Powered by AI and LLMs,” began by examining how LLMs and AI are changing communication and content at a significant speed. The digital transformation he referred to involves changes that lead to large innovations, the kind that requires participation from cross-functional teams, a clear direction, and complementary skills.

Adding that although innovation is hard it is also a priority, Goble counted Welocalize among the companies that are innovating by “embracing experimentation more than ever, unburdening teams to accept the successes (or the failures) by way of experimentation.” 

The Welocalize executive mentioned that failures in innovating often boil down to common problems, such as misaligned incentives and priorities, siloed thinking, failure to understand the customer, and lack of a process and exit criteria. 

Large Language Models as Authoring Technology

Explaining that he firmly believes that disruption creates opportunity, Goble addressed the changes brought about by LLMs and the wave of concerns raised by many language industry leaders, such as whether LLMs will be a substitute for MT, their compatibility with other technologies, and the quality of their translation or in-language output, among others. 

Goble illustrated this disruption with the example of LLMs becoming part of content authoring systems, providing the type of structure that will do away with the need to have a localization department translate content created this way. The role of the human-in-the-loop “is changing, and that human might need to go into a content authoring system instead of a traditional TMS.”

Reminding attendees that disruption is being felt at all levels of the enterprise, including CEOs, boards, and investors driving expectation for adoption, all in all changing roles, Goble added that at the same time, these changes are creating what he called “unintended opportunities for content businesses and new tech,” and urged companies to experiment to innovate and give employees the power to decide in alignment with the C-Suite.

“We fundamentally believe that humans-in-the-loop and getting humans-in-the-loop at the right place is going to be a way through the future,” added Goble.

From Strategy to Integration

Elaborating on the importance of innovation for language and content services companies at this time, Goble said that AI and LLMs provide opportunities thanks precisely “to their unparalleled ability to accelerate innovation,” and to “unprecedented access to skills, technology, resources, and more opportunities for individual contributors.”

Goble mentioned how LLMs and AI are driving a content explosion, and that they will also support some of it. He refrained from qualifying different LLMs as better or worse. Instead, Goble reiterated how the technology becomes another tool, the case of its integration into a content authoring system being an example of a solid use case.

During a subsequent interview with Slator, Goble also explained how both MT and LLMs are viable options for human translation. He compared the unknowns of current discussions surrounding LLMs and AI to those heard when NMT disrupted the industry, adding that things are not better or worse, just different.