Don’t Let ChatGPT Usher in an Era of Marketing Mediocrity!
Generative AI tools like Midjourney, and especially Large Language Models (LLMs) like ChatGPT, are such a functional leap, that they can very easily feel like magic. Don’t get me wrong, they are incredible, and have so much potential (both positive and negative). But they are not magic.
They are AI, and AI is not new. Marketers have known many of its strengths and weaknesses for years. Those don’t go away now that ChatGPT was released.
As with all AI, ChatGPT is only as good as its training data. Its training data is online content. So it’s going to produce content that is not too dissimilar from online content. In fact its whole goal is to blend in with its training data so it hides the fact that it was made by a machine.
But that’s divergent from marketers’ goals. Marketers want to stand out. ChatGPT wants to blend in.
In general, for tried-and-true methods, and for popular categories, this tool is going to be extremely useful. If you want to automate the use of best practices for writing social or blog posts, it’s a great tool. If you want to create content for a popular category with lots of history, like gardening or CRM or sports equipment, it’s extremely useful.
Generative AI is also great at idea generation, meaning it’s helpful with creativity (at least in that sense). It’ll help you come up with way more ideas for blog posts, or marketing email subject lines, or ad copy, than any single human ever could!
But only so long as those blog posts and marketing emails and ad copy are about a known subject. Just go watch ChatGPT education videos. All the examples they use are about known subjects.
领英推荐
But if you want to generate a new thought, identify a new trend, launch an innovative product or service, or uncover new data, it will be of extremely limited use.
For instance, I tried to get ChatGPT to write this article for me! I failed. It wrote similar articles, about how LLMs run the risk of marketing mediocrity, but for different reasons. It was focused on reasons that have been written about plenty by others, because that’s what it does. It said LLMs could lead to mediocre marketing because content it writes could be biased, or formulaic, or focused more on quantifiable metrics than on creativity.
That’s all true, albeit controllable with a little human oversight. But they are not the specific point I’m trying to make in this article, namely that ChatGPT and other generative AI tools are not going to offer originality and innovation.
In my opinion, that’s because this particular drawback of generative AI isn’t talked about enough. Which is why ChatGPT couldn’t write this article. Which is why I felt the article needed to be written in the first place!
So, this is a plea to marketers. Use ChatGPT and other generative AI tools for the things they’re good at. Which is a lot of stuff!
But don’t fall prey to the sea of sameness that is bound to get even worse than it already is (which is bad). Be original and innovative, and stand out. Develop new concepts and data. Be memorable and inspire change!?
But just understand you’re going to need humans to do those things.
Managing Director at Transom Group
1 年Whenever teams collaborate on marketing, sales or product copy it is an iterative and incremental process punctuated my unique insights and often outlier perspectives that illuminate. Scraping the existing internet and using AI, in my opinion, is just a form of another team member adding value to a process that in many ways is always and organically being refined.
Chief Technology Officer | UC Berkeley Alumni
1 年Jake Moskowitz, the article is a good read and the limitations mentioned "limited by the scope of its training data" are vanishing soon and its just a matter of time with the release of newer models built upon Online Neural Networks, Online Support Vector Machines (SVM), Online k-means Clustering (unsupervised learning), Online Principal Component Analysis (PCA) (live data) and Reinforcement learning algorithms, such as Q-learning and SARSA. Generative AI continues to evolve, it can acquire new knowledge and skills, much like a human student pursuing higher education (access and using historical and live data for education). Generative AI excels in Marketing and overcome Human Inefficiencies: Speed and scalability, Data-driven decision-making, Consistency and reliability, Continuous learning and adaptability, Cost-effectiveness, Personalization. In reference to relying on empathy, compassion, emotions, feeling for creating marketing content, Its just a matter of time as AI can invoke empathy (just like the Bing AI invoked Anger and Frustration) with enormous developments in sentiment analysis, and emotional intelligence in AI. This brings to "But just understand you’re going to need humans to do those things" how long ?
Principal, C.R.O. Partners, Inc.
1 年Preach, Jake. Great points.
IBM Executive, CMO, World Class Thought Leadership
1 年Well said Jake Moskowitz