by:Jean-Marie Cognet on: September 18, 2026The value of video learning is well-established. Both instructional designers and faculty rely on it for flipped classroom models—allowing students to focus on developing soft skills—or simply to diversify the media formats within an online learning path.
Numerous studies highlight the benefits of video, particularly regarding knowledge retention and self-paced learning. Students can learn whenever and wherever they want, rewatching the content as often as needed. Furthermore, the more interactive the video—prompting learners to click, explore supplementary resources, comment, or answer a quiz—the higher their satisfaction. They become actively engaged in their learning journey!
Until now, two main approaches have dominated the production of these video lectures:
- A "premium" approach, using a dedicated recording studio equipped with professional gear like high-end cameras, microphones, and green screens.
- A "pragmatic" approach, relying on personal equipment—recording directly from a computer using a webcam and an easy-to-use software solution.
In both scenarios, however, it’s a well-known fact that instructors often feel much more comfortable speaking to a live audience than staring alone into a (web)camera. This reluctance partly explains why the flipped classroom model is taking time to become the widespread standard, even though countless studies have proven its superior effectiveness compared to traditional, lecture-based teaching.
Given this reality, any technology that can simplify this "pain point" for educators is closely monitored. With recent advancements in AI, we’re seeing a surge in EdTech solutions offering to replace the instructor with an avatar. The teacher simply prepares their script, slides, or screencast, and the technology handles the rest! Add in the fact that the scriptwriting itself can be AI-assisted, and the whole package becomes incredibly tempting...
Meanwhile, a few key players have secured record-breaking funding rounds to perfect their solutions. HeyGen, founded in 2020, has already raised over $75 million from various investors in under six years. Synthesia, for its part, announced a $200 million funding round in January 2026, bringing its total investment to nearly $500 million since 2017. These resources give them the means to match their ambitions, resulting in highly polished, professional-looking videos.
But in my view, this shouldn't be an end in itself: no amount of funding can truly replace the human element.
The very foundation of education rests on humans teaching other humans. There is an inherent social connection—however subtle—between instructor and student. Allowing oneself to be replaced by an avatar in an instructional video breaks this fundamental promise; the promise that learning is more than just passive content consumption. Some educators argue that this frees up their schedule for 1-on-1 student mentoring. Personally, while I don’t dispute the estimated time savings, I’m still not entirely convinced. I firmly believe that student engagement with the course material simply cannot reach the same depths.
Getting back to the topic of time savings, I also have to question the revision process. Anyone who uses AI knows that these tools aren't entirely foolproof. At UbiCast, we experimented with these technologies—not to generate avatars, but to dub our product tutorials. On that aspect alone, correcting errors and handling the subsequent video editing ended up costing us a significant amount of time. We even found ourselves wondering if we wouldn't have been better off just embracing our French accents and recording the English tutorials ourselves!
Beyond this fundamental issue, there is the equally thorny question of Intellectual Property (IP). Using these platforms means agreeing to feed them scripts, images, slides, and screencasts—all of which constitute the core course material, and for which the instructor holds the IP rights. Even more concerning, creating an AI avatar bearing your likeness strikes me as highly risky. Most of these companies are subject to extraterritorial laws, and both HeyGen and Synthesia rely on US-based vendors. It's a slippery slope from there to finding unauthorized digital clones of yourself online, with absolutely no control over what words are put into your mouth.
Ultimately, do the risks outweigh the rewards? On one hand, you get the promise of saved time, and more importantly, faculty no longer have to face the dreaded camera lens. But on the other hand, what exactly are we giving up? The social connection, along with control over our own image and voice. Personally, I've already made my choice!
-2.png)
