On Thursday, 30 October, we looked at some websites providing tools to create AI-generated videos. There are quite a few online. Many were linked to ChatGPT so that users can use text prompts to “create” their films. We played around with generating video clips from uploaded photos.
Chris looked at: https://artlist.io Artlist
You have to register with an e-mail address first, then you can make up to four AI images and one 5- second video for free. After that, you have to pay. Chris made a mess of things at the meeting, and got lost on the site. (Tip: From the main page, click on “AI Image and video” ). She was more successful at home afterwards. Here are the results.
Here’s the original photo. It’s of Chris joyfully hugging a signpost in a forest.(We weren’t lost after all)
“This is the text prompt suggested by the programme.
I added the bit about the wild boar.”
“Animate a hiking scene where a female hiker smiles while hugging a trail signpost in a dense forest. The background features tall trees with green foliage and a dirt path winding through the woods. Add gentle wind effects causing trees to sway lightly and birds to flutter in the background. The hiker’s backpack sways slightly as she leans on the sign. Include soft sunlight filtering through the leaves, creating a warm and inviting atmosphere. A wild boar runs down the path behind her.”
“This is the resulting video. I was supposed to subscribe to download it. Instead, I used Windows Snipping Tool to take a screen recording.
You can’t see much in the way of birds or swaying trees, though the wild boar makes an entrance and the hiking stick flaps in the breeze. Next time I’ll tell it that there should be an elephant running down the path..”
We had better luck at the Club when using Veed https://www.veed.io/tools/image-to-video-ai . Again, you must create an account using an email address. Then you are given some “AI credits” and a “Team space”.
We started a “New Project” using the Image to Video option. Then uploaded a photo of JCC President, Margaret and told the AI to use a British English voice to make an animation of her saying: “Happy Halloween from the greatest witch in the Universe.“
Here’s the result.
There are many AI video generation websites and apps. The most popular one is Sora (https://openai.com/sora/) This is closely linked to ChatGPT. You can create images for free from text prompts, but (it seems) you have to subscribe to make videos. Some of the videos produced by Sora from users’ prompts can be seen here: https://sora.chatgpt.com/explore/videos
Chris showed several AI-generated videos on the Web:
The Baby Trump series: These feature Donald Trump’s real voice with an AI-generated “Baby Trump” and his baby colleagues (and enemies!) “A bad escalator and a bad teleprompter” — Trump implies UN sabotaged him, while he saved millions.
Here is the late Queen Elizabeth II of the UK enjoying herself rapping. https://youtube.com/shorts/u6ZeW1tuvMg?si=dMLt2EHTIzf4PJxA
Trigger warning! Some members thought this was disrespectful! 😉
Here is Helen Mirren shedding the years through the magic of an AI video: https://youtube.com/shorts/9P2AFIF_KD8?si=p6g7Yu2aKUrk0nAR
Photograph (of a Photograph) – Nickelback . This inventive AI video is a spoof of Canadian Rock band Nickelback’s 2005 song “Photograph”. It uses the AI-cloned voice of the singer to add stupid lyrics to the original tune while showing surreal AI-generated video. For comparison, here’s a link to the YouTube video of the original song: https://www.youtube.com/watch?v=T3rXdeOvhNE&list=RDT3rXdeOvhNE&start_radio=1
Reddit as a resource
The online forum Reddit has several sub-forums dedicated to discussions about AI, including one on AI-generated video: https://www.reddit.com/r/aivideo/. It’s worth scrolling down the Reddit page to look at some of the weird and wonderful AI creations.
Chris highlighted “Bing Chilling – Traditional Chinese Opera” on the SoraAI sub-forum of Reddit https://www.reddit.com/r/SoraAi/comments/1of0fxp/bing_chilling_traditional_chinese_opera/
mainly because the creator explains in the text what software he used, the steps he took and the source of his inspiration ( a video of American wrestler and actor, John Cena, promoting the film “Fast and Furious 9” holding an ice cream and speaking in Chinese). The AI clip shows this as a Chinese opera with the promotional lyrics sung in Chinese (see the subtitles). It is absurd, but very realistic. The song was created using the music AI app, “Suno”.
So, what can we make of all this?
Who (or what) is the creator? The AI or the person using it?
How can we tell if something is real or fake?
How can the mis-use of these AI tools (Deep fakes) be controlled?
How can intellectual property be protected ?
Is an AI actor a real actor? Is AI music real music?
There’s a lot to think about!
Christine Betterton-Jones – Knowledge Junkie