The Story Behind AI Video Tools: Hits, Misses, and Weird Glitches
A Quick Guide to What’s Promising and What Still Needs Work
These days, AI-generated videos seem to be everywhere, so I decided to take a closer look at what tools are available. Spoiler alert: overall, I came away less than impressed. Still, a few options offered some surprisingly clever features.
To keep things simple, I started with a short list and tested how far I could get using only the free versions — or whatever each tool provides at no cost. One quick note: every tool mentioned below requires creating an account with an email and password before you can get started.
Most of these tools follow a similar pattern: once you enter your email and login, you receive a small number of credits that you quickly burn through trying out the AI features — text-to-speech, text-to-video, etc. It doesn’t take long to use them up. For example, you might upload a short, single sentence and request a five-second video and — poof! Just like that, your credits are gone, and you’re prompted to upgrade to a more expensive monthly subscription.
A few programs wouldn’t even let me download my sample project without upgrading to a paid plan first. In essence, the door is only open wide enough to give you a small peek inside the room, but you can’t enter without paying.
Anyway, here’s my initial dip into the deep, swirling pool of AI-generated content…
ArtList: “Create any video you can imagine”
Claimed to be used by more than 30 million people, ArtList allows you to create “cutting-edge” AI video, images, voice-overs, music, and more. I tested it with a simple text-to-image prompt: Overhead view of a woman floating on her back in clear water with a whale swimming beneath.
The image is below.
Note that she looks like she’s flying; she’s actually floating on her stomach, and is that even a whale? I’m not sure.
I used up all my free credits on this one image (and it isn’t even a video), and it didn’t allow me to render or download a video without paying for a subscription. So, this test didn’t get any further than that.
BeFunky
BeFunky is a bit different, since it isn’t an AI video creation tool but rather a nice creative photo application. You can upload an image and apply some great effects — sketched, comic noir, crosshatched, and several more. If you’ve been in the design game for a while, you know that standard Photoshop filters can be recognized from a mile away and often can’t be used in real-world applications. BeFunky churns out nice work at a very low price. I’ve already used some of these effects in our clients’ e-learning courses to rave reviews.
Descript: “All-in-one video editing”
Descript promises “AI video editing for every kind of video” and includes several features bundled within it. Their ecosystem features custom AI avatars that don’t look as fake as typical AI talking heads, using a polished, hand-drawn style instead.
I entered some text-to-speech, and it produced a video with good voice quality and a catchy artistic-style character. While customization is a bit limited, you can tweak backgrounds, add logos, and more. Additionally, Descript offers a wide variety of characters and styles to choose from.
The user interface takes a moment to figure out, but once you get the hang of it, the process of uploading text and getting a high-quality speaking avatar rendered is quick. Descript is a winner for me, and the ILG tech team is looking to use it on the right project in the near future.
Google Flow
Another winner is Google Flow. I entered the prompt — “Rain-soaked street with detective emerging from the shadows” — and added a short blurb of text for him to speak. Within a few minutes, it created a highly realistic video with audio that looked and sounded great. Very professional quality — almost cinematic.
The caveat, like many of these tools, is that it only allows for eight seconds of video at a time, which isn’t very long. It would take considerable effort to stitch together a longer video. Still, I’m excited to see how this software evolves.
Lightricks LTX
I uploaded a static image and asked for some different poses, and LTX created them just fine. Then, I tried out the text-to-speech function and gave it a snippet of text for the person to speak. I’m not sure where it went off the rails, but what it returned to me was some sort of completely unintelligible new language — something that sounded like English played backwards and run through a dishwasher. I mean seriously bizarre.
That said, I probably won’t try this software out on a project anytime soon.
Powtoon
Powtoon ranks lowest on the list for me. Initially, I thought this might be a good alternative to Vyond, which we use quite often at ILG. According to the website, Powtoon offers multiple illustration/visual styles, pre-made video scenes that are customizable, and realistic-talking avatars — the latter of which I decided to try out.
I picked a character, gave it a small snippet of text, and hit the button, excited to see this character come to life and repeat the text back to me. And… I’m… still… waiting. The website still shows Render In Progress (this was maybe a 10-second clip), and I’ve long since used up my seven-day free trial. Ouch. It did allow me to preview the first few seconds, but the clunky speech and mouth movements of the avatar didn’t compel me to explore this tool any further.
Final Thoughts
So, there you have it. This list is just a small sampling of the growing number of AI video generation tools starting to flood the market. Overall, I’m excited to see these tools get refined; however, for the most part, many of them currently come with a lot of headaches and require significant cleanup to be of value.
It could be worthwhile to pay the monthly subscription for my two winners — Google Flow or Descript — just to explore a bit deeper and try out all the options. Stay tuned for more, as I’m sure all of this will continue to change day by day.
_______________________
Additional Resources
Can AI Develop Your Animated Video?
Can AI Develop Your E-Learning? It Sure Can Help.