A Kiwi-Indian Harvard graduate once vying for the title of New Zealand's smartest teen claims to have harnessed artificial intelligence to fully automate the process of creating video for social media.
Soumil Singh founded a platform called Unfaze.ai, which used AI to create video content for individuals and brands.
The company had received financial support from several investment funds, including Y Combinator, Hack VC, Soma Capital and Pioneer Fund.
New York-based Singh, who studied applied mathematics and computer science at Harvard, said that while the engagement rate on TikTok was 93 percent higher than other social media networks, the need to create a large volume of fresh content was often resource-intensive and cost prohibitive.
"This creates an access barrier for many New Zealand content creators, influencers and SMEs," Singh said.
"While the production of a single video could take an agency weeks to complete and cost thousands of dollars, recent advances in machine learning have now made it possible for content creators to simply 'instruct' the artificial intelligence to produce video content to meet their specifications.
"The cost of a campaign [on social media] is centred in the production of each piece of content. To produce an effective result requires multiple different versions to be tested and, even then, the video can't be reused indefinitely.
"This is one of the reasons many companies may not even have a short-form video content strategy and rely on paid media ad platforms."
Singh said New Zealand businesses on low budgets benefitted from being able to use AI to remove this obstacle.
"We believe this technology has the potential to be transformative for thousands of SMEs around the world, providing them almost unlimited access to video content which they can use to market their products," he said.
Singh said the platform had raised $5.5 million and was using the funds to further the development of the technology and support the video creation software launch in July on its existing online platform.
He said the research would bring it closer to the production of AI-generated videos that were indistinguishable from manually created content.
"While the current technology version can produce entertaining content for individual creators, a new suite of features set to launch in the coming weeks will make it the first platform to allow a company to create images of their products with multiple backgrounds using generative AI models and then convert these images into the creation of short-form video content," he said.
ISSUE
