Difficulty: ★★★☆
Impact: High
Short-term: Creates a “brand teacher”
You take the real lecture videos and replace the teacher’s face with:
•
an AI avatar
•
a cartoon-style character
•
a stylized persona your brand owns
Voice stays the same (or AI cloned).
Teacher gestures remain.
Everything becomes copyright-owned.
Companies in China already do this with 4+ hour-long videos using GPU farms.
Why this is powerful:
•
You OWN the face
•
If teacher leaves → content still fully yours
•
You can reshoot missing parts with AI recreation
•
Massive brand consistency across subjects
Potential cons:
1.
Filler words like ‘um’, ‘ah’s
2.
Student participation needs to be removed (questions posed to students may be used for rhetoric purpose but without student’s name)
3.
Explanation may not be as smooth in some parts
Improvements:
(Overcome potential cons)
1.
Clean filler words
2.
Remove students participations, e.g. 19:58~20:01
3.
Tighten explanations
(Human errors)
*Lecture Delivery Evaluation Standard
*Essentially improve explanations
1.
Polytone - if 10% monotonous and 90% polytonous, make 0% monotonous and 100% polytonous
2.
Hand gesture - render more dynamic
3.
Zoom in/out - maintain zoomed out screen b/c excessive zoom in/out is disengaging, e.g. 13:05~13:11 (better to just highlight)
4.
Non-English language spoken - render English e.g. 49:41~49:44 (미세먼지)
5.
Annotation legibility - Make more legible
6.
Annotation colour - Ideally in accordance to PPT design guidelines
7.
Content - Add examples where explanation is in sufficient


