QuickVid is taking the tech and social media world by storm because of its unique generative AI systems which automatically create short-form YouTube, Instagram, TikTok and Snapchat videos.
The AI programme can choose a background video from its library, write a script or keywords, overlay images generated by DALL-E2, and will even add a synthetic voiceover and background music from YouTube’s royalty-free music library; all by itself.
QuickVid creator, Daniel Habib, says that the application is aimed at helping creators meet the “ever-growing” demand from their fans with new creative content.
“By providing creators with tools to quickly and easily produce quality content, QuickVid helps creators increase their content output, reducing the risk of burnout,” Habib told TechCrunch.
Launched last year on December 27, the application is still early in its development, but app creator Habib promises more personalisation tools in January.
QuickVid will be able to bring together components required for an informational YouTube Short or TikTok video, including captions and avatars.
To use the application, creators will have to enter a prompt describing the subject matter of the video they wish to create. The prompt will be used to generate a script using GPT-3, and the keywords or script will be used to automatically select a background from the royalty-free stock media library Pexels.
A voiceover is added through Google Cloud’s text-to-speech API, with features to clone coming soon as well.
Even though Meta and Google have previously showcased similar AI systems that use prompts to generate videos automatically, QuickVid combines existing AI and exploits the usual repetitive, templated format of short-form videos.
According to Habib, QuickVid is “being tested and tinkered with daily” but has limitations with selecting random backgrounds that are remotely related to the topic, while DALL-E 2-generated images reflect the limited extensions of today’s text-to-image tech.
However, the app is also facing copyrights issues, since projects built by AI aren’t defined with ownership to an individual unless authentically created without AI, as stated by US Patent and Trademark Office (USPTO).
But Habib believes creators are less likely to submit their videos for patents and are leaning more towards other creators using their clips to grow their reach.
Moreover, Chat GPT-3 is not quite as reliable as an AI source generator and is commonly known to spurt out misinformation, with racist and sexist slurs. Thus, content creators will need to moderate the content they produce through the programme before posting it on social media for their audience.