Stock footage was created as a way to give people access to things they couldn’t film themselves. Today, it’s easier than ever for someone in one part of the world to find high quality video of cities and places thousands of miles away to use in their projects. But with the advancements of generative AI video, and most recently, the launch of Sora AI, the internet is buzzing with discussions on what artificial intelligence means for creators.
What are the implications of simply being able to describe a scene and have an AI generator create it for you? What are the ethical considerations? How will it impact creative roles like filmmakers and video editors? In this post, we’ll dive into recent advancements in generative AI video, its reception among creators, why it’s too early to throw away your camera gear, and why resources like stock footage won’t be going anywhere anytime soon.
The latest advancements with generative AI video technology
Artificial intelligence has quickly advanced in recent years, including video technology and tools. Video generation synthesis, which allows AI algorithms to create highly realistic videos from textual or graphic input, has been one major milestone. In January of this year, Google Research put out a paper on Lumiere, “a text-to-video diffusion model designed for synthesizing videos that portray realistic, diverse and coherent motion.” Then in February, OpenAI announced Sora, their own text-to-video model that can generate videos up to a minute long, setting a new standard for the length generative video is able to create.
On top of that, deepfake technology has made significant strides in development. This technology creates AI-generated media that has been digitally manipulated. It’s capable of manipulating video to swap out faces, change facial expressions, and even create entirely new content based off of existing video. More so than any other kind of generative AI video technology, deepfakes have been very controversial due to the different ways that they can be used to create harmful or fake content.
Reception among creators
Overall, these recent developments have been met with mixed feelings from creators. Some see them as a game-changer that will give them unprecedented efficiency and creativity when it comes to video production. Others are more skeptical about generative AI video. They have valid concerns about the potential for artificial intelligence to replace human creativity and craftsmanship.
With AI, there’s a potential future where content becomes stale and homogenized with widespread impact to professionals within the industry. Storyblocks surveyed the artists who contribute to our stock library about AI, and while they view it as a positive change, they have questions about compensation and long-term impact on their businesses.
Challenges with mass adoption
Despite recent advancements, generative AI video is facing a wide array of challenges before it becomes available as a tool to the public. Even Sora’s team has been open in acknowledging the current limitations. Some of these roadblocks include:
- Quality: As of today, AI generated video can’t match the level of detail, nuance, and emotional resonance found in video created by people. Not to mention that currently, these programs aren’t able to create sound to go along with the video they create.
- Resources: Running a generative video program requires a lot of computational resources and expertise to develop and use it, which most creators won’t have. This is also the main reason that most generative videos you see today tend to be fairly short.
- Ethics: As AI gets better at creating realistic content, the risk of misuse and exploitation increases, as we’ve seen with deepfakes. Not to mention ethical concerns with intellectual property rights and data privacy.
- Bias: AI image creation tools like Adobe’s Firefly and Google’s Gemini have been in the news recently due to offensive racial and ethnic depictions of people. This happened despite these tools being trained on different datasets, so there’s a strong possibility that the same issues will appear in generative AI video.
The role of human-created footage and resources like Storyblocks
While the excitement and discussion around generative AI video will continue to be in the spotlight for the foreseeable future, it’s important to remember that platforms providing footage created by people around the world play a crucial role today. When you download a footage clip from Storyblocks, it’s been filmed by an independent creator. When Storyblocks was launched, our goal was to support independent artists worldwide while providing affordable, accessible footage to the creators who needed it. And it’s still our goal today.
With over 2 million high-quality assets in our library, creators and businesses alike are able to find a diverse range of stock for any and every project. With one plan, you get unlimited downloads of any of our footage in HD or 4K. So you never have to compromise on production value. We always strive to keep our demand-driven library full of the content you want to see. Whether it’s through diversifying the communities shown in our footage with initiatives like Re:Stock, or prioritizing adding content that our members are searching for.
We’re also utilizing AI and machine learning to make our library work better for you. Our search incorporates OpenAI’s large language model to show you content that goes beyond your queries and helps you find the right content faster. Our library also uses an algorithm that evaluates how new content is performing every four hours and a language learning model that analyzes how members are engaging with content, making sure that you’re always seeing the best stock assets as you’re browsing. You can also find recommendations across our different stock assets to help you find the right music to pair with your footage thanks to machine learning.
The future of generative AI video
In the coming years, the impact of artificial intelligence on the creative landscape will only continue to grow. But as it does, it’s essential that the companies developing these tools and solutions focus on their customer needs. Technology shouldn’t be determining where AI goes. It should support storytellers and help them elevate their craft. Generative video has the potential to revolutionize how we create and consume content.
Realizing that potential, however, requires addressing the challenges we outlined above to create ethical, bias-free AI content. OpenAI has hinted that Sora will be released sometime this year, though they have stated that they won’t release a model that could impact the upcoming global elections in 2024. But to make sure that AI remains a force for innovation and empowerment, not a source of division and disruption, there needs to be collaboration. Creators, industry stakeholders, and policymakers have to work together to ensure the best outcome for an AI-enhanced future.
