Many companies use AI as a popular buzzword and marketing tactic. Some companies go as far as to promote their use of AI with no clear explanation of exactly how and why they have implemented machine learning models into their products. To combat this issue, we at Lionbridge have been talking to companies about their “AI-powered” tools and software to see the extent to which machine learning is actually used. In this article, we will look at real-world use cases of AI graphic design tools in the creative industry through an interview with Kevin Tey, a Product Lead at Designs.AI, and a review of two of their products.
Why Implement AI into Graphic Design Tools?
“Designs.ai is an integrated artificial intelligence-powered online creative platform under INMAGINE,” says Kevin. “Our mission is to make the creative process smarter, faster and easier for freelancers, SMEs, startups, agencies, and marketing teams.”
Design.ai’s goal for their AI-assisted tools is to make design and editing easier for professionals and amateurs alike.
Better content with more efficient turnaround times is a win-win for everyone. Furthermore, the market for AI graphic design tech is potentially huge. “According to the 2015 United Nations Conference on Trade and Development (UNCTAD) report, the global creative industry today is valued at over $509 billion,” says Kevin.
Kevin attributes the increased growth in the market over the last few years to the rising demand for creative content across all industries. “In terms of market size, the total addressable market for the creative tools segment alone is $25 billion according to Adobe.” The large market size reflects the demand for better creative tools, but making a profit is not the main reason why we as a society should implement AI into graphic design or into any other industry for that matter.
Asking the Right Questions
Too many articles talk about “greater efficiency” as a top reason for implementing AI in a variety of industries and technologies. Greater efficiency is a means to an end and not an end in itself. The real question we should be asking when it comes to AI is whether or not we are making something that will improve human life. Making things faster is great, but we must always think of the cost of doing things at a greater speed. Will we hinder the development of a designer’s own creativity if we allow AI to do the thinking for us?
On the other hand, if AI-assisted tools can improve a graphic designer’s output and improve their overall quality of life, then we should pursue them. To help answer this question, let’s take a look at the tools Designs.AI has to offer and the machine learning algorithms at work within them.
AI-Assisted Graphic Design Tools
Currently, Designs.ai has four main tools in their suite:
- Designmaker (launching soon)
- Mockupmaker (launching soon)
Furthermore, they also have four “assistive” tools:
- Color Matcher
- Font Pairer
For the purposes of this article, we will only look at Logomaker and Videomaker.
Using AI to Design Logos
To get a feel of Designs.AI’s technology, let’s look at how to use their Logomaker tool to create a logo for a brand. For example, let’s say we wanted to create a new logo for Lionbridge AI.
First, we create an account and click on their Logomaker tool. Then we enter the name of the company we want to create a logo for. After typing in Lionbridge AI, we then select our industry from a long list of 90+ categories. We selected Science & Technology as that was the closest to our service offerings.
Next, we have to choose a type of logo. The three options are icon, name, or initial. For this project, we’ve selected the name category, which means the logo will include the full name of the company and possibly a tagline in certain designs.
Next, we confirm our company name and add a slogan. 30 characters is the max length, so I couldn’t fit Lionbridge’s slogan “Breaking Barriers. Building Bridges.” Instead, I added the slogan “Building a smarter world,” to see what the designs might look like.
After that, we are shown a grid of varying sample logos and styles. They vary in font, color, design and size. We choose the ones we like and the algorithm generates logos based on our selection here and the information we inputted earlier.
For our project, we selected LIME, datatrends, and TECHLAB. Lastly, we select the color palette for the logo. Falling in line with Lionbridge’s brand, we selected the orange palette.
Onto the fun part, we must now choose from a long list of generated logos. We can choose up to 5 before moving on to the editing process. If you don’t like any of the logos generated, you can simply scroll down and click “load more”.
The final step is to edit the logos. You can change the font, size, color, arrangement, and add containers around the words.
After playing around with font types, sizes, colors and spacing, I generated the following logo in about 30 – 40 minutes, from signing up to getting acquainted with the software and starting the project:
Does Logomaker Improve the Creative Process?
In the end, the most important thing is whether or not Logomaker is more efficient than building your own logo from the ground up.
It is definitely a faster process; 30 to 40 minutes is a great turnaround time for creating a logo. However, with speed some may argue we lose quality. Designs.AI’s Logomaker tool is certainly easy to use for beginners, but that also means there is less customization available for experienced graphic designers. We can only choose from a limited list of fonts, icons, and containers, and we cannot import our own fonts or custom icons or graphics into the tool.
In the end, I could definitely recommend Logomaker to amateurs and beginners looking to make a logo on a tight budget. At just $19 USD to download your finished logo, using Logomaker on your own is significantly cheaper and faster than hiring a professional graphic designer to create a logo for you.
Therefore, logomaker could be used for things like school projects, mom and pop stores, proof-of-concept designs, and possibly early stage startups. While professional graphic designers could find inspirations in the sample logos the tool creates, they will likely feel restricted by the process due to the tool’s limitations.
Furthermore, there doesn’t appear to be any strong use of AI in the tool. Essentially, Logomaker creates logos by combining your inputs: name, slogan, industry, color palette, and style. However, this can be done with traditional computer algorithms and the tool isn’t really learning to make better designs from our input. It simply uses our input to generate logo templates that we can adjust to our liking.
Using AI to Create Videos
While the use of machine learning in Logomaker was a tad limited, we did see great things in Design.AI’s Videomaker tool. The platform is still very new, and like Logomaker, it is aimed at beginners. However, the tool creates entire videos based solely on the algorithm’s ability to understand our script.
How does Videomaker use AI?
“There is a recent trend in the AI space which involves the application of theoretical AI techniques in image processing, and natural language understanding into creative design,” says Kevin.
“An example of this application in our products is the Videomaker tool of Designs.ai. We built our proprietary NLP technologies to understand the syntactic and semantic contexts of the user’s input text in different languages, and churn out the main keywords that best represent the user’s input text. The system retrieves the best videos from the library based on the keywords. We also use the text-to-speech AI to convert text into lifelike voiceovers.”
Simply by inputting a short script, Videomaker analyzes the text for its semantic meaning. It then chooses stock video clips and compiles the video timeline using a collection of stock footage that matches the context of your script. It then automatically generates subtitles and synthetic voiceover narration synced to the video.
AI in Videomaker
While that explanation is simplified, to do all these things requires a tremendous amount of work behind the scenes. Let’s take a closer look:
- Video Classification
All of the videos in the Videomaker library would have to be correctly classified into certain categories and keywords.
- Synthetic Voices
The company compiled a library of well-trained synthetic voices. While to the trained ear they sound robotic, the intonation and pronunciation is for the most part smooth and natural.
- Natural Language Processing
They trained their NLP algorithm to match the semantics of the script to the best videos in the library classified under keywords that match the script. Furthermore, they implemented a text-to-speech system to turn the script into synthetic voice narration.
To test Videomaker out, I created a short montage video to use as the intro to a machine learning podcast. Below is the result of 30 – 40 minutes of work getting signed up and acquainted with the Videomaker system.
Does Videomaker Improve the Creative Process?
With the Videomaker tool, you can create and edit to your heart’s content without paying a dime. However, after creating a video you’re happy with, you must pay to download and remove the watermarks from each individual video. Each video costs $33 USD, with cheaper pricing available if you buy credits in bulk.
While that pricing isn’t too high, the main limitations of Videomaker lie in its lack of customization and editing. Once the video is synthesized, you can edit the voiceover, change the synthetic voice, add or remove subtitles, and switch out video clips.
However, there are no options to split clips if you want to remove parts of them. You cannot change the play speed of clips, select transition duration, or add picture overlays. After speaking to the Designs.AI team, we learned that scene trimming features are coming soon in their pipeline. Furthermore, there are a variety of animations and stickers you can add to enhance the video.
The Videomaker tool is good for beginners and people with little video editing experience who want an easy UX to make simple videos.
Videomaker definitely makes the creative process easier for beginners making video content for the first time, and those with tighter budgets. With Videomaker, you don’t need to hire voice actors for narration, record your own footage, or create your own animations. All you need is a typed script. For advanced content creators, Videomaker shows promise once more editing controls are added.
Can AI Improve Graphic Design?
Is it possible for machine learning technology to improve graphic design and make the lives of creatives easier? Based on what I saw from Designs.AI, I think the answer is a definite yes but with a few caveats to consider.
Logomaker is a quick and easy to use tool for beginners to quickly create multiple logo designs. However, you are limited to the fonts, frames, and graphics within the Logomaker library. Professional graphic designers may feel limited by the amount of customization available.
Videomaker employs multiple AI technologies including synthetic voices, text to speech, and natural language understanding models. The impressive library of stock videos categorized based on content made it easy for the algorithm to to match videos to our script. However, it must be said that this tool is also more aimed at beginners.
For more advanced video editing, you would have to export the video you created in Videomaker to another tool and then do more detailed editing there. However, Designs.ai have created strong foundations with Videomaker.
Once the library expands, the models get stronger, and more editing functions are added, Videomaker could definitely be a go-to software for content creators. In the future, it would be amazing if content creators could simply type in their script, generate a video, do a bit of editing, and publish.
The applications of AI in the graphic design industry are vast and promising. For beginners, Logomaker and Videomaker are viable tools to quickly create logos and videos with no specialized experience.
All in all, we were really happy to see that the “AI” in Designs.AI wasn’t just a buzzword. The company does employ various machine learning models into their tools to make the user experience better for their customers. Hopefully, more companies will learn to implement ML in innovative ways to improve and enhance the tools that graphic designers and content creators use in their day-to-day work.
If you’re looking for more reading on use-cases of AI in various industries, check out the related resources and subscribe to our newsletter below. Also, please be sure to check out our Essential Guide to Training Data.