Evaluating 7 AI video generators while waiting for OpenAI’s Sora

2024-08-21 22:00:08

A lot has changed over the past year while I’ve been waiting for “the best AI video generators for creators and marketers” to emerge.

The list of leading contenders got scrambled six months ago when OpenAI announced Sora, “an AI model that can create realistic and imaginative scenes from text instructions,” on Feb. 15.

Ten days later, I tried to check out Sora for myself. But OpenAI’s website told me, “At this time, we don’t have a timeline or additional details to share on Sora’s broader public availability.”

I’ve been waiting for more updates. And waiting. But waiting for Sora is starting to resemble “Waiting for Godot.” (If you haven’t seen Samuel Beckett’s play, then spoiler alert: Godot sends word that he’ll be arriving tomorrow, but he never appears.)

The long wait: Sora’s elusive public availability

I watched “OpenAI Sora’s first short film – ‘Air Head,’ created by shy kids,” when it debuted on March 26. And I assumed that Sora would become publicly available in a few months.

But it’s already August and only a limited number of visual artists, designers and filmmakers have been granted access to “gain feedback on how to advance the model to be the most helpful for creative professionals.”

In other words, I still can’t evaluate Sora for myself.

In addition, shy kids is a Toronto-based production company. It’s unclear if KaraVideo.AI, a video generator powered by Sora, deserves most of the credit for this eye-catching video. If you don’t have a creative team of humans driving the text-to-video model, then would you see similar or different results?

Exploring 7 alternative AI video generators

While waiting for Sora, I taught two online classes on influencer marketing and AI at the New Media Academy in Dubai, UAE. I showed my students “Air Head” but emphasized that AI video generators still need human creativity to produce interesting results.

To help them apply what they learned, I assigned them the task of creating a video using a free AI generator available in the Middle East and North Africa (MENA) region within two or three days.

To find the best tools, I turned to AI search engines (Perplexity, ChatGPT-4, Claude 3, Microsoft CoPilot and Google’s Gemini). Each provided a list of recommendations, with options ranging from six to eight suggestions.

Overall, my students learned about seven free AI video generators – six that I selected and one that another instructor in the program recommended. Here’s the list in alphabetical order:

1. Canva’s Magic Studio: Quick and customizable video creation

Canva’s AI video generator is part of its Magic Studio, a suite of AI-powered tools designed to help teams quickly create video from text, and then refine the content using editing tools, applying filters or removing backgrounds with a single click.

2. InVideo AI: Simple and efficient video production

InVideo AI provides pre-made templates and an AI-driven text-to-video feature. It allows you to create professional-looking videos quickly. The free plan includes watermarked videos and limited export options.

3. Kapwing: Collaborative editing with AI tools

Kapwing is a collaborative video editor that offers AI tools for video creation, such as automatic subtitling and meme generators. The free plan has some limitations, like watermarked exports and limited cloud storage.

4. Lumen5: Turning text into engaging videos

Lumen5 converts text into engaging video content. You can input your blog post, article or any text content, and it will generate a video based on it. The free plan includes basic features, though it adds a watermark.

5. Pictory: Condensing long-form content into short videos

Pictory allows you to transform long-form content, such as blogs and articles, into short, shareable videos.

It uses AI to highlight key points and create a video summary. The free version comes with some limitations but is functional for basic needs.

6. Runway ML: Experimenting with creative AI

Runway ML is a unique tool that offers a free plan with credits for using their AI features, including video generation. It’s a good choice if you want to experiment with a creative and powerful tool, but keep in mind the free tier has limitations.

7. Synthesia: AI-driven multilingual video content

Synthesia is an AI video generator that transforms text into speech in a few minutes. It includes more than 160 AI avatars that can speak your text in over 130 languages. This enables marketers to create training, how-to, and product marketing videos for global audiences.

Insights from hands-on use of AI video generators

Following my overview of these seven AI video generators, I asked my students to pick just one to use for their capstone assignment. I also disclosed that I was writing this article and allowed my students to contribute their thoughts on the option that they had used.

Dana Al Khulaifi, a student at United Arab Emirates University (UAEU) and a Data and Technical Support Manager at Dawahi AlAin, a community project, said:

“It was easy and enjoyable to create videos with InVideo.io! My thoughts were given to the AI, which created a masterpiece out of them. I was able to produce professional-looking videos quickly because the AI handled the majority of the work.”
“I was really pleased with the range of layouts and how simple it was to customize it. It took hardly any effort to make my videos appear fantastic. InVideo.io is an amazing resource for anyone looking to create amazing videos fast and simply.”

Three other students, who wanted to remain anonymous, shared their comments about other AI video generators that they had used:

One student explored Synthesia for creating AI-generated videos. She was impressed by the availability of different Arabic dialects but noted that the pronunciation in Arabic was not perfect. She suggested that AI still needs human guidance for quality control, such as improving pronunciation and adding emotional expressions.
Another student used Runway for video creation, finding the process straightforward with tools for editing, transitions and effects. She noted that while AI could automate many aspects, it still lacked the originality and emotional depth that human creativity provides. She emphasized the importance of human oversight in bringing contextual awareness and cultural sensitivity to AI-generated videos.
A third student used Canva to create videos quickly. She noted the efficiency and speed of AI tools but highlighted the lack of emotional expression and the computer-like tone of AI-generated content. She suggested that AI tools could be used for drafting videos, which could later be refined with better AI tools or human intervention for a more polished final product.

Here is a list of comments about using AI videos from more than a dozen other students:

“I tried to use AI tools, but the video seemed very boring and didn’t really showcase anything social.”
“Comparing the AI video with the one I created, the AI video was less engaging.”
“AI is great but lacks human touch, especially in terms of storytelling and video quality.”
“The AI didn’t really match the style I wanted, and the video felt robotic.”
“The AI video was interesting but not as detailed or creative as the one created by the human team.”
“AI-generated content still needs human input; the video didn’t capture the creativity I was looking for.”
“I tried to challenge the AI, but it struggled to properly depict what I had in mind, especially with cultural representation.”
“The AI couldn’t get the hands right – it looked weird, and the voice-over was robotic.”
“AI tools provided options and brainstorming ideas but didn’t produce a final result that was up to par.”
“AI video was very different from what I expected, especially when it came to hands and backgrounds. It wasn’t fully accurate.”
“AI is good for generating basic content, but it lacks the precision and personalization needed for luxury brands.”
“The AI had issues with details like hands, and the background didn’t match the context I intended.”
“AI videos often don’t have the same emotional connection or creative flow as those made by humans.”
“The backgrounds created by AI often don’t fit the context I want.”
“AI videos can create something quickly but lack in creativity and detail.”
“It didn’t feel authentic; AI cannot replace the human touch and emotional intelligence.”
“The hands and background in the AI video were distorted and didn’t match my original idea.”

Advantages and challenges of AI in video creation

Overall, here are the general observations made by dozens of students in the two classes:

Advantages

Efficiency and speed: AI tools can quickly generate videos, providing a starting point or draft for content creation.
Ease of use: Many students found AI video tools straightforward, enabling them to create videos without extensive technical knowledge.
Versatility: Tools like Synthesia offer diverse options, such as different languages and dialects, enhancing accessibility.

Limitations and challenges

Need for human intervention: AI-generated content often requires human oversight to ensure cultural relevance, emotional depth, and accuracy.
Pronunciation and expression: AI-generated voices, particularly in non-English languages, may lack proper pronunciation and emotional expressiveness.
Content authenticity: AI tools can struggle with originality and creativity, emphasizing the need for human creativity to add unique elements.

Testing the top AI video generators while waiting for Sora

While waiting for Sora, evaluating other AI video generators can still teach us some important lessons.

The most important lesson is this: “AI won’t take your job, it’s somebody using AI that will take your job.”

The economist, Richard Baldwin, was the first person to say this at the 2023 World Economic Forum’s Growth Summit, and variations of this phrase have been repeated as people discuss the potential impacts of AI.

Don’t wait for the perfect AI video generator to finally appear.

Evaluate some of the best ones that are already available to learn what they can – and can’t – do already. Taking this test-and-learn approach will enable you to evaluate Sora when it appears.

And if Sora doesn’t arrive in the foreseeable future, then you will have learned another important lesson: “Perfect is the enemy of the good.”

Either way, you are less likely to lose your job because you don’t know when or how to use AI to “generate videos up to a minute long while maintaining visual quality and adherence to the user’s prompt,” which is what Sora keeps promising to do.