There is a new Best Evidence in Brief, and this time, I picked this very relevant study from this biweekly newsletter, this piece written by Feifei Wang, The Centre for Information Technology in Education, The University of Hong Kong.

In the age of generative artificial intelligence (AI), can generative AI-made teaching videos deliver learning outcomes comparable to those of human-made videos? A recent study explored this question by comparing the effectiveness of four generative AI-made videos with four human-made videos on the same topics in a management course. The online intervention involved  447 US-based laypeople, most of whom had completed higher education. The study included two experimental conditions:  in the first, 213 participants watched four teaching videos in a human-AI-human-AI sequence; in the second, 234 participants watched them  in an AI-human-AI-human sequence. After each video, participants completed a survey rating their learning experience, followed by a  multiple-choice exam to assess learning outcomes.

Results showed that human-made videos provided a small but statistically significant advantage in learning experience, suggesting that the participants still preferred human teachers in the videos. Regarding learning outcomes, participants in both conditions achieved a similar degree of acquired knowledge. The authors noted that as generative AI-made teaching videos become increasingly common due to their ease of production, researchers and educators should critically explore ways to use generative AI-made content to improve learning experiences, while being mindful of potential limitations and ethical considerations in production and use.