Beyond her own department, Bondi’s memo encourages all other federal agencies to follow suit, advising them to scale back the production of multilingual documents and consider English-only services. The memo also suggests agencies use AI translation to communicate with LEP individuals as a way to save money, providing a potential boost to language technology platforms.
Contrary to the Trump administration’s mandate, some states are launching new operational and legislative initiatives to make sure LEP residents receive adequate language services, including New York and Massachusetts.
In light of what could be revenue-hitting consequences for some language solutions integrators (LSIs) in the US, we asked readers what they thought about Bondi’s memo, and the majority (55.9%) think it is worse than expected. A third (30.2%) think the compliance memo is as expected, one in nine (11.6%) see it as better than expected, and the rest (2.3%) have no opinion on the matter.
How Good is AI Translation?
The language industry entered Q3 2025 with AI translation available wherever there is a cloud — and fast becoming part of a growing number of devices as live AI translation-as-a-feature (TaaF).
There is no shortage of AI translation offerings from known and new language technology platforms (LTPs), including multilingual video and audio, live speech translation, and AI accessibility.
Alongside major announcements from big tech, innovations from new language AI companies continue to pop up. The startups are increasingly differentiating themselves by becoming vertical specialists, offering comprehensive solutions for specific sectors like healthcare, gaming, and marketing, instead of just providing core language functions.
With so many offerings, it is now possible to compare and contrast many different AI translation modalities and providers, so we asked readers if they think AI translation has gotten better in the past 12 months. Almost one in two respondents (42.0%) believe it has definitely improved. A little over a third (36.0%) think it is not the case, and the rest (22.0%) think it has improved somewhat.
Bubble or Unparalleled Opportunities?
Perhaps it is too early to tell if, in macroeconomic terms, the AI momentum is an echo of the dot.com bubble. Regardless, when tech bubbles burst, there tends to be a relatively large talent pool left over for those who prevail, but while momentum is high, as is the case in 2025, competition for AI-skilled workers is nothing but fierce.
At the time of writing, Meta has been making news by snatching famous AI researchers for its so-called “Superintelligence” group with equally super compensation packages. That includes researchers from Google, ScaleAI, and OpenAI.
In fact, OpenAI CEO Sam Altman claimed in June 2025 that Meta was offering USD 100m signing bonuses for researchers to switch companies. At least five went for it.
Slator 2025 AI Dubbing Report
The 85-page report analyzes the supply and demand for AI dubbing and the technical and operational nuances in delivering AI dubbing across verticals.
There is no question that teams made up of the most innovative and influential minds in AI can give Meta an edge. Where things are not as clear is what superintelligence means to the tech giant in material terms.
Case in point, now that agentic AI effectively begins to replace entire task-driven teams and signs of industry consolidation ring loud, will the big fish, e.g., Meta, end up with some sort of elite thinkers and tinkers at the expense of promising startups?
We asked readers if they would be keen on joining a language AI startup. Most (43.6%) would not, thinking it too risky. Close to one in two (41.0%) would do it if the product is great, and the rest (15.4%) are already working for an AI startup.
Let the AI Tell It
That is exactly what a lot of us are doing: asking ChatGPT, Gemini, or another chatbot to analyze, summarize, translate, and read aloud all the TLDR text it can handle that we cannot, will not.
AI can also dub, which involves all of the above but the summarization function, as well as transcription.
Specifically, AI dubbing takes speech-to-text (STT) to transcribe and then translate the original message, proceeds next with text-to-speech (TTS) to [ideally] generate a realistic and expressive synthetic voice that captures the emotion and nuance of the original performance, and moves on to the obvious final result: a translated and natural-sounding synthetic voice.
To be convincing, the AI dubbing technology must also ensure accurate lip synchronization, where the new audio aligns seamlessly with the on-screen mouth movements. All these components must work together in a finely tuned and complex process that looks easy on an AI app’s interface.
One or two short videos are not much of a chore with any of the latest generation AI dubbing apps. Creating multilingual AI dubbed feature-length videos on a large scale, on the other hand, requires sophisticated management of multiple video and audio tracks for distribution.
In those scenarios, which carry a higher level of risk, a dynamic marketplace is emerging, where various players, including language technology platforms (LTPs), language solutions integrators (LSIs), and large enterprises, are all competing for a piece of the action.
We asked readers how often they watch AI-dubbed content, and the vast majority of respondents (77.0%) said they never do. About one in 10 do so sometimes, and two equal-size groups either do so often or rarely (6.3% each).
