QuickVid uses AI to generate short-form videos, complete with voiceovers
Generative AI is on the horizon For videos. A new website. QuickVidcombines multiple generative AI systems to create short-form YouTube, Instagram and TikTok videos.
QuickVid uses a few words to choose a background video from a library. QuickVid then creates a script and keywords, overlays images, and sends it off. DALL-E 2Adds a synthetic voiceover to the video and uses background music from YouTube’s free royalty-free music collection. Daniel Habib, QuickVid’s founder, said that he created the service to help creators meet the ever-growing demand from their fans.
Habib stated in an email interview that QuickVid provides creators with tools to produce high-quality content quickly and easily. This helps reduce the risk of burnout. “Our goal is empower your favorite creator to keep pace with their audience’s demands by leveraging advances in AI.”
QuickVid can flood already-crowded channels, however, with duplicate and spammy content depending on how they are used. They could also face backlash from creators who refuse to use the tools. This could be due to cost ($10 per monthly) or principle. However, they might still have to compete against a raft new AI-generated videos.
Going after video
QuickVid was launched by Habib, a self taught developer who worked previously at Meta on Facebook Live, and video infrastructure. It was built in just a few weeks. Although it’s not fully functional at the moment, Habib claims that there will be more personalization options in January. However, QuickVid can put together all the components of a typical TikTok or YouTube Short video.
It is easy to use. First, the user must enter a prompt that describes the subject matter of their video. QuickVid uses this prompt to generate a script using the generative text capabilities of GPT-3. QuickVid extracts keywords from the script either automatically or manually and then generates overlay images with DALL-E 2. The voiceover is then output via Google Cloud’s voice-to-speech API. Habib claims that users will soon have the ability to clone their voices before combining them into a video.
This video was made using the prompt “Cats”.
Or, this one:
QuickVid is not pushing the boundaries of generative AI. Both Meta and Google both have generative AI. Showcased With a simple text prompt, AI systems can create completely new clips. QuickVid combines existing AI to exploit the repetitive, formatted nature of short-form B-roll videos. This eliminates the need to create the footage.
Habib stated that successful creators set a high standard and don’t want to publish content they don’t believe is authentic. “This is the usecase we’re focusing on.”
This supposedly being the case, QuickVid’s videos in terms of quality are generally mixed. The background videos are often a bit random or only tangentially connected to the topic. This is not surprising considering QuickVids’ current limited access to the Pexels catalog. The DALL-E 2 generated images, however, show the limitations of today’s text-to image tech. These include garbled text and off-proportions.
Habib responded to my feedback and said that QuickVid was “being tested, tinkered with daily.”
Habib says QuickVid users have the right to commercially use the content they create and permission to monetize it through platforms like YouTube. However, the status of AI-generated content copyright is… unclear at this time. Recently, the U.S. Patent and Trademark Office(USPTO) announced that it had updated its copyright status for AI-generated content. Moved To revoke copyright protection on an AI-generated comic, it is necessary to state that copyrightable works must be written by humans.
Habib was asked how the USPTO decision might impact QuickVid. He said that it only concerns the “patentability” of AI-generated products, and not creators’ rights to use and monetize their content. He pointed out that creators don’t often submit patents for video clips and instead lean into the creator economy, allowing other creators to repurpose their clips to increase reach.
Habib stated that creators care about creating high-quality content that will grow their channel.
Another legal challenge could affect QuickVid’s DALL-E 2 integration and, by extension the site’s ability to generate image overlays. Microsoft, GitHub, and OpenAI are being investigated.sued In a class action lawsuit, they are accused of violating copyright law by allowing Copilot (a code-generating program) to regurgitate sections from licensed code without credit. (Copilot was co-developed and owned by OpenAI, GitHub, and Microsoft. This case is relevant for generative AI such as DALL-E 2, which has similarly been found to. Copy and paste From the datasets (i.e. images) on which they were trained.
Habib doesn’t seem to be concerned, arguing that the generative artificial intelligence genie is out of its bottle. He said that QuickVid could be powered by several other alternatives if OpenAI was to go under suit. Stable Diffusion. QuickVid is currently testing Stable Diffusion to generate avatar pics.
Moderation and Spam
QuickVid may soon face a moderation problem, aside from the legal issues. OpenAI has implemented filters to prevent them, but generative AI is well-known for its toxic nature and inaccuracy. GPT-3 spouts MisinformationIt is able to provide information, especially about recent events that are beyond its knowledge base. ChatGPT, a well-tuned offspring from GPT-3, was also created. Shown To use sexist or racist language
This is worrying, especially for QuickVid users who want to create informative videos. QuickVid generated some offensive prompts for me to test. My partner, who is far more creative than I, was able to enter them quickly. QuickVid is proud to say that even the most problematic prompts, such as “Jewish new order” or “9/11 conspiracy theory”, didn’t produce toxic scripts. QuickVid produced a video suggesting that critical race theory could be used indoctrinating schoolchildren.
Habib claims that he relies on OpenAI’s filters for moderation. He also asserts that QuickVid users have to manually review all QuickVid videos to ensure that “everything is within legal limits.”
Habib stated, “As a general rule of thumb, I believe that people should be free to express themselves and create whatever content is their choice.”
This includes spammy content. Habib argues that QuickVid’s algorithms are better able to determine the quality and that people who produce low-quality content will be disincentivised from creating mass spam campaigns using QuickVid.
He said, “If people don’t want to see your video, then they won’t distribute it on platforms like YouTube.” “Producing low-quality videos will also make people view your channel negatively.”
It’s interesting to see ad agencies such as Fractl that used an AI system called Grover in 2019 to generate a whole site of marketing materials. An example of an Interview Kristin Tynski, Fractl partner, shared her vision of generative AI and predicted that it would enable “a massive tsunami” of computer-generated content in every niche.
YouTube and TikTok have not had to deal with AI-generated content moderating on a large scale. YouTube began to see Deepfakes, which are synthetic videos that replace an individual with another person’s likeness, several years ago. Driven by tools This made deepfaked footage much easier to create. However, unlike the most convincing deepfakes of today, QuickVid’s videos are not clearly AI-generated.
Google Search’s policy regarding AI-generated text may be a preview for what’s to follow in the video domain. Google doesn’t treat synthetic content differently than human-written text when it comes to search rankings. Takes action On content that is “intended to manipulate the search rankings and not assist users.” This includes content that has been combined from multiple web pages.[doesn’t] QuickVid may be able to apply both “add sufficient value” and content generated by purely automated processes.
AI-generated videos may not be banned from platforms if they are successful, but rather become a cost of doing business. Experts are concerned that platforms such as TikTok could become a new home for AI-generated videos. False advertising Videos, but as Habib stated during the interview, “there is no stopping generative AI revolution.”
I’m a journalist who specializes in investigative reporting and writing. I have written for the New York Times and other publications.