How I Actually Use AI at Ecolinguist (and Why It Matters)

Ecolinguist is... a lot. It’s a channel, it’s a show, it’s an educational ecosystem. And it’s hard to explain quickly — especially to people unfamiliar with ideas like mutual intelligibility or conversation-based learning.

I use AI tools to shape my ideas into messages that actually land — whether it’s writing a pitch, updating the homepage, or just organizing my thoughts.

Before AI? I was often silent. I didn’t know how to talk about the project clearly, and that made it hard to reach out for help or explain the potential of the platform. I could make a video — but describe what I do and why? Total block.

Tool used: ChatGPT (outlining, adapting tone, drafting ideas)

Let me tell you — writing good website copy is no joke. Even as a language teacher, I found it painfully difficult. It would take me days to write a single blog post, and often I’d give up halfway through. Meanwhile, I knew how important it was. A confusing site could turn people away before they even understood what Ecolinguist offers.

AI now helps me generate structure, headlines, and drafts that I can refine. It’s not “click to publish” — I still spend a lot of time rewriting — but the difference is I’m no longer stuck at the starting line. And honestly, that’s what I needed: a co-writer who doesn’t get tired.

Tool used: ChatGPT (for structure, copy tone adjustments, and iterative rewriting)

If you’re familiar with the Ecolinguist channel or the Academy, you know how much I value listening input — but also how rare it is to find natural, level-appropriate, and relevant listening material. That kind of content has always been hard to produce at scale.

On the YouTube channel, I’ve worked with amazing volunteers to record language challenges — and I still do. Real voices are irreplaceable. But technical issues, mic quality, and scheduling often slowed things down. In the meantime, I’ve started experimenting with AI-generated challenges too. They still need proofreading and tweaking, but they’ve helped me keep the format going when human recordings aren’t possible.

For the Conversation Labs and listening exercises, AI voice tools finally made it possible to produce consistent, natural-sounding audio in multiple languages — fast. And I only started using them when the quality finally felt human enough. Earlier versions were… truly awful. I waited a long time for this to become viable — because anything less would’ve just made the learning experience worse.

Before this? Authentic content was too advanced. Textbook audio was limited, expensive, and not adaptable. I was stuck. Now I’m finally creating the kind of input that helps learners move forward.

Behind the scenes, I’m also building a system to make it easier for more real voices to be included — sustainably and at scale — so that community members can contribute when they’re ready.

I’m genuinely excited about what text-to-speech technology means, not just for my work, but for the whole language learning community. It opens up endless possibilities for the future of language learning. It’s a massive shift in how learning materials can be created — faster, more personalized, and more aligned with how people actually learn. It’s not just a helpful tool — it’s a game changer. Especially for minority languages with scarce audio input resources.

Tool used: Realistic Text-to-Speech Generator — e.g. ElevenLabs

You might not realize how much work goes into subtitles — especially multilingual ones. I’ve had videos sit unpublished for months (or even years) simply because I couldn’t get good enough captions.

Transcriptions used to take ages. English translations were often rough, because not all volunteers were fluent in English. And when you’re working with lesser-known languages, there’s no auto-caption option. I couldn’t even check the subtitles in languages I don’t speak.

Now AI tools help speed up transcription and translation — but we still do a manual review, especially when captions are burned into the video and can’t be changed later. It’s not perfect, but it’s manageable. And for many people, those captions are what make the content usable. Some researchers working with minority languages have even asked to reuse the transcriptions for their work. That’s how big a difference it makes.

Tools used: Whisper (for transcriptions), DeepL (for translation), experimenting with custom scripts for subtitle file handling

In the early days of the Ecolinguist show, every new volunteer got a 1-on-1 Zoom call with me. I’d check their audio, explain the format, and give mini media training. It was lovely — and completely unsustainable.

I started using AI to help me write training materials, onboarding messages, and quick adjustments for different roles (e.g. guest vs host). It doesn’t replace personal interaction, but it makes the admin side of organizing community projects so much lighter. Now I can scale what used to be one-on-one to many people — and still make it feel personal.

Tool used: ChatGPT (for customizing message templates, simplifying complex info)

This is the kind of thing no one talks about but takes SO much time: copying text from PDFs or screenshots, fixing broken formatting, cleaning up subtitles, restructuring layouts…

AI helps me do this mechanical cleanup work in seconds instead of hours. That means I can stay focused on teaching and creative work instead of getting sucked into formatting purgatory.

Tools used: ChatGPT (for text cleaning and reformatting)

I avoid using AI to generate culturally specific images (because it’s not reliable yet), but I do use it to help with quick visual fixes — like removing backgrounds from images, resizing things, erasing clutter, or generating simple illustrations for concepts I can’t find in stock photos.

I’ve also tried using AI-generated images in language learning materials — but stopped, since odd proportions, extra fingers, and general visual glitches were more distracting than helpful.

Without this, I’d either spend too long in Canva or Photoshop — or just give up and go without. This way, I get what I need to make materials clean and clear, without going off track.

Tools used: Canva Magic tools (for resizing, background removal), DALL·E (for basic concept illustrations when stock fails)

Many of the people who appear in my videos — especially in the language challenges — are recording from home, using basic equipment or in noisy environments. That used to be a big problem. Even if the content was great, poor audio could make it unusable.

Now I use AI-based audio enhancement tools to clean up recordings: reducing background noise, improving clarity, and making sure voices are easier to understand. It’s not magic, and it still requires some manual work — but it often saves recordings that would have been unusable before.

Tool used: AI Audio Enhancement (e.g., Adobe Podcast, Auphonic, etc.).

If you’ve never tried to combine, convert, and translate SRT files, you’re lucky. It’s messy, frustrating, and error-prone — especially across languages.

AI tools now help me manage subtitle workflows faster and more reliably. This used to be one of the biggest bottlenecks in publishing episodes. Now it’s doable — and that alone feels like a superpower.

Tools used: Subtitle Edit + AI-assisted scripts, Whisper for transcription, DeepL for quick translations

Honestly? The biggest gift AI gave me is momentum. Before, I was constantly switching between writing, translating, teaching, editing, emailing, organizing, onboarding — no wonder I was burning out.

Now AI takes care of the mechanical, repetitive tasks, so I can focus on the stuff that really matters — designing lessons, engaging with learners, improving the platform, and recording new shows.

It doesn’t make the work easy. But it does make it possible.

Tool used: ChatGPT (my daily assistant and sounding board)

If you’ve followed Ecolinguist for a while, you probably care about people, languages, nuance, and cultural depth. And you might be wondering — how does all this AI fit into that?

Here’s the thing: I don’t use AI to replace people. I use it to fill gaps that were slowing me down, burning me out, or blocking the project from growing. I still proofread. I still teach. I still design every learning path and video idea myself.

AI doesn’t write my lessons or make my videos. It helps me get to them faster.

I believe that with the right balance — transparency, thoughtfulness, and a deep respect for language — AI can support, not replace, human connection. That’s the whole point of Ecolinguist.

So if you’ve ever wondered how this all gets made — now you know.

And if you’ve been one of the people supporting this work, or cheering it on quietly in the background: thank you. You’re why I keep going — AI is just how I keep up.

I’m fully aware that AI raises serious concerns — from misinformation and copyright issues to surveillance, bias, deepfakes, and the growing wave of low-quality content (“AI slop”). These issues are already affecting many industries in very real ways, and as an individual creator, I don’t yet have clear answers on how to solve them.

But I’m paying attention. We’re all navigating this moment together — as a society, and as humans. AI is here to stay, and I’m trying to use it with care, purpose, and transparency — not to replace people, but to support meaningful work.

My focus is on finding ways to use these tools to improve lives — especially to expand what’s possible in language education.

What’s your relationship with AI in your work or learning? I’d love to hear.