Voice cloning for education: It is exactly like this picture
This is part of a new series within our blog wherein we will be exploring the educational applications of different pieces of Artificial Intelligence technology.
Voice Cloning technology like that offered by 11labs has been in the headlines quite a bit in the past few months. Nefarious use cases like the cloning of prominent individuals for political or fraud purposes have dominated discussion about the technology. These abuses exist and need to be taken seriously, but the fact a tool can be used irresponsibly does not mean it is by nature bad.
Voice cloning technology is a process by which an AI model receives audio recordings of the subject and uses those to infer what the voice would sound like pronouncing different words. The cloned voice can then be adjusted for volume, speed, excitement, or any number of other variables. A few months ago as a proof of concept I cloned my voice in about 20 minutes using 11labs platform and then produced this video wherein ChatGPT 4.0 produced a script which was then fed to Gamma to produce a slide deck and to 11labs to produce a voiceover. In the video linked above the only thing I do is advance the slide deck. In the examples below I am assuming you are an instructor and you have used the technology to clone your own voice.
First, it is a way to diversify your content delivery for students. You have probably produced an incredible volume of written work for your courses like assignment guides, syllabi, and examples. Perhaps you even assign a book or article you have written for your class. With a voice clone you can offer a different mode of delivery with very little effort. Got an article? Now you have an article and a podcast. This has profound implications for workload and accessibility. Of course, text-to-voice technology has existed for some time, but with the intimacy of using your own voice your students can feel connected to you and the content in a different way. This is a common use-case for NLW who hosts the daily podcast AI Breakdown (which I strongly recommend). He has also used the tool to create content when his voice is out due to illness. I think we all have concerns about abuse with this kind of use. If you leave an institution, but they paid for your access to the technology, can they still use your voice and likeness as part of a course? These are messy questions with a direct analog of existing video, so while these problems are complicated, they are not new and your institution may already have policies.
Second, voice cloning technology is a way to reach new audiences where they are at. One of my great life regrets is that I do not know a 2nd language. Voice cloning technology allows the user to deliver the same content in a wide range of languages. This means you can maintain the intimacy of your own voice and meet students closer to where they are at. This is not the same as knowing other languages, but for someone like me the path from where I am at to producing alternative language assignment descriptions or vignette videos explaining concepts is years of study and practice. Now that path looks more like a few clicks and some vetting of the content from a multi-lingual colleague. Again, this points towards using technology to meet students where they are and is an exciting application of the technology. Abuse is of course possible here too, as is clumsy implementation wherein a translation is done in an offensive or inaccurate manner. The argument is not that the tools are perfect or without defect, but that they open up new possibilities for instruction.
Broadly, the use of voice cloning allows instructors to maintain or even increase the connection they have with students with minimal investment of time. Training the programs does not take long and the results are quite good. While writing this I kept feeling like I was missing a lot and as is often the case with new technology applications I probably am. I look forward to hearing about good and clumsy implementation of this technology from colleagues as we all find our way in this exciting, yet uncertain, future.
There are ongoing developments we highlight on our blog. We also recommend the blog of Ethan Mollick. We look forward to hearing from and working with you as these developments unfold.