Day 9: AI-generated deepfakes
Welcome to Day 9! Today we’ll be exploring deepfakes and thinking about their creative potential in education, as well as the challenges they present in distinguishing real from fake material.
Today’s post is authored by Ruth Powell, Head of Digital Learning Practice, UAL.
—
If you’re an active user of social media, you may be familiar with apps that allow you to swap faces of people in photos or change a photo of you to show what you might look like in 50 years’ time. Many of these apps use AI in a technique termed ‘deepfake’.
Whilst these types of application might seem to be a bit of fun, you’ve probably also come across more worrying media reports of deepfakes used maliciously to exploit or mislead people; spreading misinformation without the direct consent of the people involved.
You may hear of deepfakes being referred to as ‘Synthetic Media’, as illustrated in the following definition:
What are deepfakes?
A deepfake is an emergent type of synthetic media that uses artificial intelligence and machine learning (AI/ML) techniques to create or manipulate audio, video, images, or text. Deepfakes produce highly realistic and convincing content of events or people doing or saying things that never actually occurred.
The “deep” in deepfake refers to deep learning, a type of machine learning that uses neural networks to generate fake images, videos, or audio recordings. The term “deepfake” is a combination of “deep learning” and “fake”.
As this technology evolves, creating deepfake content will become easier to do. Detecting what is real and what isn’t is likely to become an increasingly difficult challenge. Understanding this technology will help us respond effectively to its challenges whilst leveraging its potential.
Uses of deepfakes in creative arts industry
Filmmaking is one industry where deepfake technology is being actively researched and adopted. Studios use AI-based techniques to achieve special effects, such as ageing characters, or for dubbing characters in other languages as in this video where David Beckham appears to deliver his malaria awareness message in nine languages.
Uses of deepfakes for teaching materials.
There are various ways to practically use deepfakes in teaching materials, for example developing synthetic media processes with a combination of AI tools to create teaching materials. There are also various considerations to take before using deepfake technology in teaching.
Read these articles to explore further:
- Why we need to talk about deepfakes (Leon Furze)
- Deepfakes are coming for higher education. What can we do to prepare? (Times Higher Education)
- Uncovering Deepfakes: Classroom Guide + Discussion Questions (AI for Education)
Activity 1
How good are you at detecting what is real and what is fake? Test your skills by playing this Real or Fake game.
Activity 2
Have a go at creating your own AI Avatar using HeyGen (account creation needed). HeyGen is an online platform that uses AI to create videos, including AI-generated avatars and voiceovers. It comes with several AI Avatars that you can use, mitigating any risks from recording and uploading a sample video of yourself.
- Prompt ChatGPT to create a short (approx. 2 mins) narrative on a topic of your choice.
- Create an account and log into HeyGen.
- Select Create video / Avatar video and choose the orientation.
- Select the avatar you’d like to use.
- From the menu down the left-hand side, select Script.
- Copy and paste your script from ChatGPT into the script box.
- Select Enter under the script box.
- Select Submit.
It will take a few minutes for the video to be processed.
Alternative tools
These are all free to use but you are required to create an account.
Discussion:
Join us in the Teams space to share a link to your AI avatar in the Teams space and contribute your thoughts to the questions below to the discussion. If you don’t have access to the space, email us at teachingexchange@arts.ac.uk for the attention of Hannah
- How do terms like “deepfake” and “synthetic media” shape our perception of AI? Does the name influence your sense of trust and does it matter?
- Can you see any potential for synthetic media to enhance the way you deliver educational materials or create engaging learning experiences?
- What ethical considerations or risks might arise from using deepfake or synthetic media in an educational setting?
- What practices would you put in place to balance the creative potential of deepfake technology with the need for academic integrity and authenticity?