Your professor keeps their face, their voice and their name.

If you have sat through a demo of an AI video tool in the last two years, you have watched a synthetic presenter read a script. Synthesia raised $200 million at a $4 billion valuation in October 2025 on the strength of it. HeyGen and Colossyan sell the same idea. Colossyan markets a 90% production cost saving. Panopto bought the avatar company Elai in 2024 and now sells generated video as an add-on to the repository your lectures already sit in.

The pitch is compelling on a slide: type the script, pick a presenter, get a video. No camera, no schedule, no faculty member who hates being filmed.

Then you take it to the academic council.

What actually happens in that room

Someone asks whose likeness it is. Someone else asks whether a student paying two lakh rupees is entitled to be taught by a person. A senior professor asks, more quietly, what exactly the university is proposing to do with thirty years of their teaching. And the head of the faculty association, who has been waiting for this, asks what happens to the teaching post once the avatar can deliver the module.

These are not unreasonable questions and there are no good answers to them, because the tool is doing precisely what it says: replacing the lecturer.

We are not that

Lecture Studio starts from a recording of a person teaching and returns a recording of that same person teaching, better presented. Specifically:

This is not a moral position we adopted to be comfortable. It is a product boundary, and it is the reason the conversation with your senate is short.

The research does not favour the avatar either

There is a widespread assumption that a presenter on screen improves learning. The largest study on the question — Kizilcec, Bailenson and Gomez, across two field studies covering 12,468 learners — found that adding the instructor's face raised perceived social presence and cognitive load, but that "learning outcomes and attrition did not differ." Thirty-five percent of learners chose to switch the face off entirely, citing distraction.

So the case for a synthetic presenter is not that it teaches better. Nobody has shown that. The case is that it is cheaper to produce than filming a human — which is a real advantage if you have no footage. If you already have hundreds of hours of faculty recordings, you have already paid that cost, and generating a substitute throws it away.

Where a synthetic presenter genuinely makes sense

We would rather say this than pretend the category has no use. Generated video is a reasonable answer when there is no recording and no realistic prospect of one: a policy update that must go out in four languages by Friday, a compliance module nobody wants to present, a course whose author left the institution two years ago. That is a real job and the tools do it well.

It is a poor answer to "our lecture videos are not good enough to publish," because the problem there is not the person in the frame.

What we ask for, and why

None of that is regulatory theatre. It is what turns a faculty meeting from an obstacle into a formality — and it is an answer the avatar vendors structurally cannot give, because the thing you would be consenting to is the replacement itself.

Sources