Last week, I moderated a webinar on AI Adoption in Interpreting Workflows, organised by GALA’s Special Interest Group Interpreting. The aim was modest but necessary: to look at how AI interpreting is actually being used today, and to invite an open discussion among practitioners and stakeholders. To do that, we brought together two viewpoints that rarely meet: Dell Technologies, a global corporation using AI interpreting internally, and Boostlingo, a service provider offering both human and AI solutions.
What emerged was a clear, if still early, picture of where things are heading.
The Numbers from Dell
Dell’s presentation was striking precisely because it avoided speculation. Instead, we saw internal data describing their multilingual services today:
- Over 500% growth in AI interpreting usage.
- Nearly 50% reduction in costs.
- High satisfaction levels—about 80% of users reporting a positive experience.
- AI used both to expand access and, in some cases, in place of professional interpreters.
These figures complicate the common narrative that AI interpreting spreads only because buyers are misinformed or dazzled by marketing. Those factors exist—but they do not fully explain what is happening. In many contexts, the technology now performs well enough, expectations are low or practical, and the economics are difficult to ignore.
And importantly: we are still on the upward slope of improvement. Even without breakthroughs, current systems will likely continue to advance in the coming years.
Boostlingo and the Education Sector
Boostlingo, for its part, shared insights from U.S. education—a sector where multilingual needs keep rising even as language-access policies remain inconsistent. Their forecast suggests increasing reliance on AI interpreting simply because the demand outstrips available human resources. As with Dell, this is less about ideology than about constraints and practical needs.
A Trend, and a Lack of Attention
The examples discussed in the webinar are not broadly generalisable. But they offer valuable signals—early indicators of how organisations are adapting multilingual communication to new technological realities. Understanding such signals is essential for researchers, professional associations, and policymakers.
And yet, participation reflected a different reality: only one academic tuned in, and, as far as I could tell, no representatives of professional associations. This absence is troubling. Decisions about the future of interpreting are being made in offices, classrooms, and corporate workflows, and too often without the people who should be helping shape that future.
Looking Ahead
None of this means human interpreting is disappearing. But the role it plays is changing, and the pace of that change is accelerating. By 2030 — again, a conservative estimate of mine — AI interpreting may well become the default option for many everyday multilingual situations, even as professionals remain indispensable in more complex, high-stakes settings.
If we want to understand and influence this shift, we need to pay attention. And we need to be in the room when these conversations happen. The entire presentation (not the discussion that followed) is available on the GALA Resource Center with many more insights, opinions and interpretations than what I present here.