We can /imagine how in the future there will be more to presenting ideas, visions, & scenarios than Powerpoint.
The problem nowadays is that it takes a designer to be able to put together more complex designs, illustrations, and 3D models. Will Generative AI change that?
Creatives, AI & Visual Media
The creative field is one where AI is gaining a foothold, and in this article we look more into where and why that traction is taking place.
According to Salesforce, roughly 62% of Marketers are using Generative AI to generate images, photographs, etc.
An array of applications are hitting the market to aid in the creation of images, videos, 3D visuals, and everything in between.
Midjourney is a popular text-to-image GenAI application that has experienced explosive growth since its inception. As of October 2023, Midjourney had hit the $200M in ARR (Annual Recurring Revenue) mark.
in April, Guo and Chenlin Meng, a fellow Ph.D. student, dropped out of Stanford to launch Pika to build an easier-to-use AI video generator. Since then, some 500,000 people have tried the software and it’s now being used to create millions of new videos each week.Forbes
Luma AI is company that enables users to create text-to-3D GenAI models. The company just raised $45 Million to scale operations.
In each category mentioned above – images, videos, 3D – there are a plethora of other companies competing in the space. Capital is flowing into these companies partially because demand is so high that they need to continually purchase GPU (Graphic Processing Unit) power from the chipmakers.
The AI boom is most visible in the stock market where companies like NVIDIA have seen their share price explode multi-X over the last years. But outside of the big VC investors and the GenAI entrepreneurs themselves, the best way to dive into the creative side of the applications is to start /prompting.
Prompts, Generation, Augmentation
What are Prompts?
It all starts with a /prompt. There is no software to download, and in most cases these applications offer some form of a ‘free trial’ to become familiar with the application.
Prompts are like command- line ‘directions’ to tell the software what to produce. The level of depth that can be entered into an application like Midjourney is huge.
/imagine is the Midjourney prompt to describe to the AI engine the type of photo desired.
ie. /imagine a midsummer’s dream.
Despite the fact it is ‘AI,’ there is an entire lexicon and structure to the commands of an application like Midjourney. There are several other types of commands, which can be learned on YouTube Channels like Theoretically Media.
A person could spend hours learning about just the commands, but for Midjourney Prompts in particular, here are some of the additional Prompts:
- /describe image-to-text command to see how to describe certain images
- /panning allows the image to stretched in multiple different directions
- –parameters — allow for instructions about aspect ratio, other image parameters
Each application has its own Prompts, which is why the educator circuit on social media, YouTube, etc. is so important.
Unlike some of the AI Article Generators for writing – where articles are produced based on relatively simple commands or keywords – knowledge about certain technical photography terms or other relevant domain expertise will help to produce better results.
Theoretically Media has the bases covered for how to use Midjourney, from beginner to advanced.
Once the Prompt is made and the GPUs process it, the Visual Media will be generated. A ~4s video on Pika will take about 30s.
Pika Labs will ‘Generate’ short videos from literal text, ones that can be downloaded, shared (subject to commercial terms), and ultimately edited.
From looking at various examples of creative projects using Midjourney, Pika, or Luma (among hundreds of other GenAI applications), it is apparent that many Creatives will take an asset (image, video, 3D) generated in one of the above applications and use a variety of typical softwares in that field (ie. Photoshop) to create a final output.
A lot of skill is required to turn these generated ‘assets’ into any type of immersive visual media.
The business model of GenAI’s is a Saas-style, monthly subscription model.
Pika for example, has four pricing tiers that offer various benefits in addition to monthly usage tokens.
Theoretically Media lays out how to use Pika Labs like a Pro.
Looking at the media produced from all of the above-mentioned applications (including Luma AI), a lot of it is very raw and rarely ‘fit for purpose.’
Meaning that the production from these applications, in their current form, needs to be edited, tweaked, and ultimately augmented.
They ultimately still require Professionals to turn them into the desired creative assets, whether it be for business or personal projects.
These types of GenAI apps in the Visual Media category have enormous potential to revolutionize fields related to creative media, but they are still raw and in their infancy.
Trust in Visual Media Made by AI
In our post on Sales, AI, and Business Writing, it was mentioned that consumers are already getting sick of being inundated with AI-generated content. There are already issues with trust:
Only 26% of consumers trust content created by AI (May 2023), while 38% neither trust nor distrust it, with 36% outright distrusting it according to Hubspot.Sales, AI & Business Writing
Visual Media is a different field with different inputs/outputs compared to written content. There are tonnes of new applications that are hybrid marketing studios, turning ‘AI assets’ into Ads, Marketing Content, and Sales Material.
But a lot of the outputs from Visual Media are very raw and suited towards more creative expression than corporate branding. Furthermore, many artists (at least at this juncture) cite the various GenAI tools used to produce their works.
53% of Marketers believe these GenAI tools are a “gamechanger” according to SalesForce, so we know these tools are being used as much in business as for pleasure; there will be ethical concerns with the outputs.
On the other hand, once mastered, these applications allow ‘non Creatives’ to experiment with new types of visual media and come up with fun ways to express their ideas.
A lot of what these applications produce lends itself to the world of Sci Fi at this juncture, but who knows, it may only be a matter of time before someone comes up with a creative way to express an idea and uses these new Visual Media tools to fund their own venture.