Raphael Thys
The generative future, made by author, on Midjourney
I donāt think anyone could have predicted the year we had with Generative AI in 2022. If we continue at this pace, we are in for a wild ride in 2023.
The following is my humble attempt at some predictions for Generative AI in 2023. Here are five that I believe will happen.
1. Corporate Tools Will Integrate Creative AI At Scale
While Microsoft is currently integrating DALL-E2 with their Designer and Image Creator applications, I suspect that Microsoft will also leverage other OpenAI models such as GPT-3 for their office suite of tools.
Considering Microsoftās investment partnership with OpenAI, Itās likely that Microsoft adds AI writing features to MS Word and also leverage OpenAIās whisper for speech recognition in MS Teams video-conferencing.
During Adobe Max 2022, Adobe featured a plethora of new AI features as part of their creative cloud suite, but expect more generative AI assistance in all of Adobeās apps ā not just for editing, but also for content creation, such as text-to-image in Photoshop and text-to-video in Premiere.
2. AI Generative Video Matures For Production
Just like AI art and imagery have matured in 2022 (cough cough Midjourney v4), we will see the same type of maturity for AI generative video.
image excerpt from Googleās 2022 AI@event
Google has already showcased its Imagen+Phenaki video model and while it is still a bit rudimentary, I would expect it to rapidly increase in quality and performance.
Metaās Make-A-Video and rumors of an official Stable Diffusion text-to-video tool will hit public release in 2023, and we should expect some fantastic creative video ventures from video content producers.
Basically, 2023 will be the year of AI Generative video in many forms ā for video shorts, storyboarding concepts, music videos, Tik Tok video memes, animations, etc...
3. The Prompt Is The New Search
In 2023, we will see new and interesting ways that prompting will be used.
A current example is how Metaphor.systems uses prompting ā which is utilizing a form of link prediction ā resembling how someone on the internet might refer to a link. check it out.
Composing in a text field will become more common than simply searching. An example might be when stock imagery is requiredā¦ I can search for ābears in a lakeā and hope to get back some royalty-free results to choose from, or I can simply prompt itā¦ and get newly composed image(s).
Synthesis engines will eventually emerge as the de-facto engine for internet user input. 2023 will be the start of that transition.
This article goes into more depth on this idea.
4. Media Agencies Will Embrace Generative AI
I know this is happening already, but I think 2023 will be a breakout year when AI generative content will be integrated as part of creative business workflows and processes.
Ad agencies, marketing firms, publishing firms, design, and creative service agencies, etcā¦ will all be utilizing AI generative art for final product releases, and also for brainstorming, inspiration, and drafts.
Generative text and imagery will be used in ad and marketing displays, posters, magazines, online articles, and campaignsā¦ really for almost anything.
AI generative music will also make its way into jingles, online content, podcasts, and various forms of media production.
Just to throw in some preliminary examples of text-to-design, the following are some magazine covers I experimented withā¦ no editing was done here, this is all from text promptsā¦
5. Generative AI Copyright and AI Ethics Concerns Will Rise
In case you havenāt noticed, there is quite a heated debate on the ethical concerns of using AI-trained artistic content. There are also concerns about bias in datasets and the lack of multicultural representation.
As more and more generative content goes mainstream, there will be greater demand for more regulation and policy. As of today, there is no clarity on who owns the copyright on generative art, itās a murky unregulated art form at the moment.
Iām not sure any of this will be resolved in 2023, but it will certainly take center stage.
In Summary
2023 will be an explosive year for generative AI.
I predict some of the main highlights will include enhanced corporate tooling, AI generative video, New and creative prompting applications, Media companies using AI content, and an escalated concern for copyright and ethical AI.
Let me know if I missed something.