AI Content Generation made a splash in 2020 and 2021, and 2022 will likely be no different.
AI Content Generation is the automation of language achieved by training AI models on a huge corpus of text. Developers can ask these AI models to continue writing where they leave off. The approach is to iteratively estimate the word that most likely comes next given the context that came before it.
Several companies brought their enormous AI language models to market and enabled a plethora of startups to build products around the most advanced publicly-available AI models to date. This was largely due to a breakthrough in methodology in 2018 and generous rounds of funding like this one ($1B from Microsoft) and this one to train these big models.
These AI platforms are analogous to the large horizontal cloud computing platforms that power virtually every new software application. With access to a broad set of underlying capabilities, developers are free to create niche applications that solve real problems for the rest of us.
(Before we get too far, we just want to thank you for checking us out! We share insights on AI for marketing professionals with updates on the latest technology and tools. To keep in touch feel free to email Andrew at firstname.lastname@example.org or join our new LinkedIn Group.)
This new era of widely available AI has led to a commoditized technology virtually overnight. The barrier to entry to building an application with a bare bones UI powered by sophisticated AI has dropped to essentially zero. This is seemingly great for everyone!
- The platform companies get to rake in per-token-generated fees from developers.
- Developers get to build on top of technology they didn’t have access to the year before
- Endusers get to experience real-world applications of AI that hopefully solves meaningful problems for them.
A Low Barrier Entry is a Blessing and a Curse
Similar to other commoditized services like cloud computing, a low barrier to entry leads to many people with the same ideas building the same thing very quickly. This is what we saw (and are seeing) with AI-powered marketing applications.
If you want a taste, check out the products on ProductHunt that mention GPT-3 here.
The AI language generation platforms are so good that there really doesn’t require much for a developer to ship an MVP product with AI as it’s backbone. The obvious problem with this is that the market gets extremely muddied and it’s hard to sift through the myriad of products that pop up overnight.
The products that will end up winning are those that lean on other core competencies and do not rely solely on the power of these AI platforms.
AI-Powered Content Tools are Better Partners than Replacements
Despite the promising new wave of AI-enabled products, there seems to be some hesitancy in the content creation world about “AI taking writing jobs.” That isn’t going to happen any time soon, if ever. AI systems do a great job helping writers and they do a bad job at replacing writers. There is not yet a replacement for true human creativity and the ability to understand the emotional needs and intent of customers.
There are certainly some cases where you can let an AI system loose on a given topic and the output is logically consistent and scarily close to something an actual writer would produce. However, in the vast majority of cases, the output is still a hodgepodge of logically consistent text blocks stuck together.
These models get us 90% of the way there, and the last 10% is going to be a very very difficult hurdle to jump. The challenge immediately in front of us is how do we use this solution that gets us “90% of the way there” to our advantage?
Luckily, the commoditized nature of the technology is putting a very powerful tool in the hands of an army of capable developers who are hopefully doing the hard work to find out what marketers need. The application layer on top of the commoditized AI layer is what is going to provide the real differentiated value between all of these products.
After all, it wasn’t a commoditized cloud service* that brought us “Stranger Things” and “Squid Game”, it was the platform** built on top of it.