How AI Is Transforming Localisation for Live Content
AI/ML made swift subtitling de rigueur on YouTube and for videoconferencing several years ago. But now LLMs are changing the game for real-time translation and localisation and language-mapped reanimation for news and other live content, as PADEM Media Group’s Allan McLennan, Alchemy Creations’ Andy Beach, and Dubformer’s Anton Dvorkovich discuss in this clip from Streaming Media Connect 2025.
Improvements in Cultural Localisation
Allan McLennan, founder and president of PADEM Media Group, asks Andy Beach, founder and principal of Alchemy Creations, to discuss the evolution of the streaming industry’s approach to connecting cultures via subtitling. “Are we starting to look at something new and different when it comes to connecting live and news into new forms of localisation, new types of interaction? Are you sensing some of that?”
“Absolutely,” Beach says. “I think the tech has continued to mature and the latency has gotten to a point where even AI can help assist on the live feeds so that if you’re in a scale position or a scale role where you’re managing a lot of different channels, you can actually leverage it.” He references an earlier conversation when they talked about how much “better and crisper” cultural localisation has gotten, to the point that it can offer regional variations within languages.
The localisation can now apply to particular generations and ages. “I had mentioned earlier the notion that my 82-year-old grandmother-in-law and my 15-year-old daughter are not going to want to watch the same version of something because one of them’s going to be lost in the language, but they can technically get their own versions at this point that are a little more specific to them,” Beach notes. “And we’ve even got AI that will help do better lip sync along the way as part of it.”
How AI Aids Live Captioning
McLennan wonders where AI learns how to do this interpretation. Beach answers that there’s always a generalised large language model attached, but “the best models or the best direction that I’ve seen for, particularly, live captioning, is when you can front-load a lot of the additional source. So if it’s sport that you’re talking about, make sure you’ve got all the names in there—the team names, maybe the oddball pronunciations of an athlete’s name that might not be common. Preload all of that into some sort of augmented retrieval so that you’ve got that in there.” He adds that the same advice applies when localising for cities: “The exact same spelling’s going to be said three different ways across the country at some point, and you want to make sure that you’re saying it right for your area that you’re in regardless of it.”
These “pieces don’t have to be part of the LLM—they can be a secondary data source that you go retrieve from and pull from,” Beach notes, adding that this option is faster, “so that you’re not waiting on a big model to be updated in order to take advantage of it.”
Building Emotional Engagement Through Dubbing and Voiceovers
“I’m just going to roll right over to Anton [Dvorkovich] because you’ve been doing this worldwide in multiple regions, geographics, areas, languages,” McLennan says. “Is that what you’ve been interpreting?”
Dvorkovich, CEO and founder of Dubformer, replies that his company focuses on dubbing and voiceovers, which he sees as growing in demand. “Our view and our mission is about creating content that can really connect with audience[s] on [an] emotional level,” he states. “We’ve been seeing several experiments and nicely set up tests where, for example, voiceover was compared to subtitling in terms of audience engagement, and we see that it does create better engagement. So this is where we’re focusing our attention: trying to build something emotional.”
Join us 24-26 February 2026 for more thought leadership, actionable insights, and lively debate at Streaming Media Connect 2026! Registration is open!
Related Articles
The now-familiar narrative of livestreaming in the pandemic tells us that when social distancing compelled streamers to migrate from on-prem to remote production and the cloud, massive cost savings were our reward. But cloud-based streaming also carries hidden costs. NBCU's Paul Kirchberg, Live X's Corey Benkhe, and Sargeway's Sarge Sargent discuss the true costs and savings of cloud and REMI operations and how to avoid the pitfalls in this clip from Streaming Media Connect 2025.
07 Jan 2026