AI-generated content is everywhere now, and this includes video production. "AI" is a broad and loosely defined term, and it encompasses a number of different technologies. AI is used in other tools featured on this site, for example, to generate closed captions, or correct mispronunciations in audio.
For our purposes, we're discussing services that allow you to create an AI avatar or persona, feed it a script, and generate video content from scratch. There are a number of these services, and the specific companies, features, and limitations are shifting all the time.
Before we discuss the options that are available, it's worth considering when AI use makes sense for teaching.
The obvious appeal to these services is they take a process that can be time-consuming and hard to learn--recording yourself--and allow a machine to do the work for you. This can be a huge benefit! But just like using video in general, it's important to think through whether the tools actually serve your learning objectives.
Consider some of the primary reasons to record video in your teaching:
To offer asynchronous content outside of class meetups for "flipped" classrooms or hybrid and online courses.
To cover content that especially benefits from visuals, like worked examples, software demonstrations, or lessons based around visual frameworks.
To offer faculty presence and foster a sense of connection with your students.
Let's look at how AI video applies to each of these needs:
AI videos can certainly generate asynchronous content. However, just because the video is easier to produce, consider whether it is the best medium. If you had to record the video in a traditional way, would you still produce it in this format?
If material needs accompanying visuals, then AI generation can expedite some portions of the process like creating narration. But you'll still need to combine that narration with the appropriate slides or visuals, with correct timing. Many AI video platforms allow for this, but it still requires time on your part-- the entire video cannot be automated.
An automated video, even if it bears your likeness, is unlikely to foster a sense of connection with students. At present, these services still bear some hallmarks of artificiality. Students will know it's AI. And if they didn't, that would pose serious concerns about transparency.
Consider the value the video adds to your student's experience. If all you're doing is automatically generating a narrated version of text content, then it is, at best, a convenience feature. And students can replicate that functionality on their own with tools like Speechify, ElevenReader, or NaturalReader.
AI may still not be the best idea for creating complete teaching videos from scratch. Where it can be incredibly helpful is as a part of a larger workflow.
Using a cloned voice, for example, might allow you to correct an error or mispronunciation in a video without going back to re-record it. Or you may be able to use AI to generate narration for a video but pair that narration with slides and visuals you've created more carefully on your own. One notable platform in this regard is Descript, which is featured elsewhere in this guide. Descript is still structured around recording your own content, but features many helpful AI-enabled tools to speed up recording and editing.
If you are going to generate AI video, consider whether duplicating your own likeness is the best option. You may want to consider creating a "Virtual TA," in order to maintain a clear separation between content that features you and content that uses AI material.
However you approach using these materials, transparency with your students is vital. Be sure to communicate where and how you're using AI-generated content. This not only helps maintain trust with your students. It models the behavior you expect from them as they complete coursework and assignments.
If you feel you have a use for AI generated video, the excerpt below describes some of the most popular tools. This video has been AI-generated, itself, in order to provide you with an example of the results.
Because of the computational resources required to generate and host video, these tools are paid services. Some of them may have free trials, but they will require a subscription if you want to generate a significant amount of content.
Some of these tools are geared towards content generated at scale by larger teams. Synthesia and HeyGen stand out as having accessible options for individuals.
Synthesia - Synthesia is one of the leading platforms in the AI video generation space. You can create a digital twin of yourself by uploading a brief video, then generate videos in a variety of languages with lifelike avatars that deliver your script. It’s widely used for corporate training and educational content because it offers a user‐friendly interface along with a large library of premade avatars and voice options.
HeyGen - HeyGen is another popular tool—especially if you want more control over creating your own avatar and using your own voice. It allows you to generate a custom AI avatar from a short video clip or even a photograph. Users appreciate HeyGen’s flexibility for personalizing the avatar’s look and voice, which can be a great asset when producing teaching videos that need a personal touch.
Elai.io - Elai.io offers multiple approaches for creating your digital twin. You can choose among several avatar types (such as “selfie,” “studio,” or “photo” avatars) based on the kind of video you want to produce. This platform is tailored for educational and corporate communications, making it a strong choice if you need to produce a series of teaching videos quickly while maintaining a consistent on-screen presence.
Colossyan - Colossyan is geared particularly toward educational and training content. It lets you either use premade avatars or create a custom one through a simple recording process. In addition to avatar generation, Colossyan also offers interactive features (like quizzes) and script-generation tools to help you craft engaging and informative lessons.
AI is a constantly developing field. We expect the capabilities and limitations of these tools to change significantly over the coming months and years. Though the exact path this technology will take is unknown, we recommend regularly checking in on the technological landscape, regardless of whether or not you elect to use these platforms.
Students and educators are not only navigating rapid advances in what these tools can do and the quality of their output, but shifting cultural norms about the acceptability of their use. Bear in mind that these tools are available to your students, just as they are to you, and just like with AI use in general, you may need to have an open and direct discussion about whether they are permitted, and if so how, in completing student assignments.