I have been awestruck by the capabilities of tools like HeyGen and Synthesia in the way they can create videos voiced by AI avatars and, with HeyGen in particular, translate from one language to another. The latest beta tool in HeyGen enables someone with limited technical skills (i.e. me) to create an AI avatar of themselves. This is a screen recording on me conversing with my twin. I could choose to speak with it/ him in a number of languages and on topics outside the grounding though it is bland and vague in those spaces. Apparently with some nudging away from the highbrow his Hindi is pretty good and his French sounds Quebecois. I grounded it in the King’s central guidance on GenAI and a few other things I have written. For now only sharing the the recording while I get to grips with the implications of this. I have honed the base prompt so that the twin is crisper in response and doesn’t waffle on. What do you think of this? What are genuine educational use cases that are not about putting humans out of work?