Warning: this content is older than 365 days. It may be out of date and no longer relevant.

One Step Closer to the Marketing Singularity

We’re one small step closer to the marketing singularity, the event where machines become our first choice for doing marketing work. Ever since OpenAI’s announcement of GPT-3 (and the relatively heavy restrictions on it), a number of other organizations have been working to make alternative models and software available that have similar performance.

As background, GPT-3 is the latest in the family of transformers, machine learning models that can generate text and perform exceptional recognition of language. These models are large and very computationally-intensive, but they’re also generating text content at quality levels approaching human. GPT stands for Generative Pre-trained Transformer, and they’re becoming more accessible and powerful every day.

Let’s look at an example, using EleutherAI’s GPT-J-6B model. Let’s take a relatively low-value marketing task like the drafting of a press release. I’ll use this release from a plumbing company:

Page 1 of release

With the text shown on screen only, I fed it to GPT-J-6B. Let’s see what it came up with:

Synthetic release

And for comparison, here’s the rest of the original release:

Original release page 2

I would argue that what the machine synthesized is easier to read, more informative, and generally better than what the original release presented. More and more AI-based tools will hit the market in some form that are at least “first draft” quality, if not final draft quality. We’ve seen a massive explosion in the capabilities of these tools over the last few years, and there’s no reason to think that pace will slow down.

So, what does this mean for us as marketers?

I’ve said for a while that we are moving away from being musicians to being conductors of the orchestra. As more easy and low-value tasks are picked up by machines, we need to change how we approach marketing from doing marketing to managing marketing. These examples demonstrate that we don’t necessarily need to hand craft an individual piece of writing, but we do need to supervise, edit, and tune the outputs for exactly our purposes.

In terms of your marketing technology and marketing operations strategy, you should be doing two things.

  1. Prepare for a future where you are the conductor of the orchestra. Take a hard look at your staffing and the capabilities of the people on your team, and start mapping out professional development roadmaps for them that will incorporate more and better AI tools for easy marketing tasks. Those folks who aren’t willing to invest in themselves and pivot what marketing means are folks that you might need to eventually transition out of your organization.
  2. Be actively testing and watching the content creation AI space, especially around transformer-based models. Everything from Google’s BERT, LaMDA, and MUM models to natural language generation to video and image generation is growing at accelerating rates. Don’t get caught by surprise when a sea change occurs in the marketing technology market space – by being an early adopter and tester of all these different tools and technologies, you’ll be ahead of the curve – and ahead of your competitors.

Tools like the GPT family are how we will execute more and more of the day to day tasks in marketing. Prepare yourself for them, master them, and you’ll be a marketer who delivers exponential value to your organization and customers.


You might also enjoy:


Want to read more like this from Christopher Penn? Get updates here:

subscribe to my newsletter here


AI for Marketers Book
Get your copy of AI For Marketers

Analytics for Marketers Discussion Group
Join my Analytics for Marketers Slack Group!