When ChatGPT took the world by storm in 2022, many wondered if it and other forms of generative AI would spell the end of the need for writers. As writers, naturally, we took notice. Our conclusion so far? We don’t think these systems are coming for our jobs just yet. But they are here to stay. In this white paper, we look at how companies can use generative AI to support their marketing plans, examine some of the limitations the technology still faces and consider what our relationship with these tools might look like. We have therefore prepared this white paper series in three parts:

Part 1 (below) was written by the writers at Ascribe using prompts from leadership.

Part 2 is the exact same white paper, with the exact same prompts that were provided to our writers, except given to ChatGPT.

Part 3 is a comparison of our work to that of ChatGPT.

What will be discussed

1. What is generative AI?
2. How generative AI can support marketing
3. When not to use generative AI
4. Treat AI-generated content as a first draft
5. How we can work with generative AI

What is generative AI?

Generative AI made headlines in 2022 when OpenAI released ChatGPT, a chatbot built on a foundational large language model. ChatGPT and similar generative AI systems use massive data sets and natural language processing to create new text in response to user-entered prompts. These systems represent a massive leap forward from previous AI-type text generators because of their ability to interpret and fulfil requests written in the way people talk. So for example, while an older system might depend on keywords or inputs provided in a specific formulation, newer generative AI systems can take a request like “give me a simple one-paragraph explanation of quantum physics” and deliver exactly what was asked for.

Understandably, ChatGPT and other systems like it have sparked a lot of conversation about what the future of content creation and consumption will look like — and what role humans will play in it. Although the potential for disruption is huge, we believe it will be a long while before there’s no need for human involvement in content creation. In the meantime, there are plenty of ways to work with the technology instead of being scared of it.

How generative AI can support marketing

Anyone who’s had the opportunity to experiment with generative AI has likely considered how it could be used to make their job easier or to develop content at a lower cost. And while we wouldn’t recommend using it for long-form content like blog posts, reports or white papers (see the next section for why), it can be useful for shorter, simpler pieces — especially for a small business with a limited marketing budget. Generative AI can also support some of the preliminary stages of content development, leaving more time to focus on the actual writing.

Here are a few ways to use it effectively:

• Idea generation: If you’re not sure exactly what to write about, you can ask it to suggest some topics relevant to your field and your audience.

• Outline development: Once you know what you want to write about and have a list of the main points you want to make, generative AI can help you organize your thoughts into a logical flow.

• SEO optimization: If you have an existing piece of content you want to optimize for SEO, generative AI can help integrate relevant keywords.

• Social media posts: When you want to promote existing content on your social media channels, generative AI can summarize the piece and create engaging posts according to each platform’s requirements.

When not to use generative AI

While generative AI can be useful for some kinds of content, it’s not a perfect system. And the longer the content piece, the more its flaws become evident. Even if you use it only for short pieces or support tasks, it has some critical limitations that make it unsuitable for some purposes.

We recommend not using generative AI in these five situations:

1. You want to demonstrate thought leadership.
   Generative AI doesn’t actually think for itself, but rather pulls ideas from other content that’s already out there. It might put those ideas together in novel ways, but ultimately, it can’t say anything that hasn’t been said before. So if you want to share thought leadership, you’re going to have to produce it yourself.
2. Your content relies on current data or events.
   ChatGPT’s large language model only includes content published up to 2021. That means anything it produces won’t reflect the most recent knowledge or developments. And although some generative AI systems pull their content from online sources and may be more up to date, there’s still no way to vouch for the accuracy or currency of the sources they’re using.
3. Your content needs to be locally relevant.
   Even if you ask the generative AI to include sources to back up any claims, it may not recognize the nuances of your specific market. For example, content about heating and cooling systems can vary significantly across different climate zones, so the information the system pulls might be accurate for the southern United States but completely irrelevant for readers in northern Canada.
4. You need to retain copyright of your content.
   Who owns the content you publish? If it was written by generative AI, the answer may not be clear. Furthermore, some of the content the model was trained on could itself be copyrighted, which could put you in a legally sticky situation if the original copyright holder objects.
5. Your company has a distinct brand voice and style.
   Although generative AI can vary the tone, style and voice of the content it produces to a certain extent, it’s ultimately limited by the data set it was trained on. That means it’s unlikely to ever be able to produce something that truly matches your brand voice.

Treat AI-generated content as a first draft

No matter how you decide to use generative AI, it’s never a good idea to take what the system produces and just publish it without reading it over. Treat anything produced by generative AI as a first draft, rather than final copy. Your edits might be as simple as tweaking the wording to better reflect your brand voice. Or they could involve fixing glaring errors that could be highly embarrassing to your company. OpenAI itself acknowledges that its model can sometimes produce text that is mechanically and logically sound but completely inaccurate.

Another good reason to review all AI-generated content is that AI models and the algorithms that inform them often show unintended biases. While you might think an algorithm wouldn’t be prone to the same sort of biases humans are, the fact is, these systems were built by humans and trained on data sets generated and selected by humans. And because their capabilities are largely based on pattern recognition, they’re extremely good at finding patterns — even ones humans might have missed. Those patterns often reinforce existing biases, and programmers have yet to figure out how to mitigate this issue. That makes it critical for you to review content carefully to ensure it’s not implying something you didn’t intend.

A note on cost: While cost might seem like a good reason to use AI-generated content (it costs essentially nothing for instant delivery of as much content as you want), it’s only truly cost-effective if the results are close to exactly what you want. For a short, simple piece that just needs some minor adjustments, it might be worth considering. But the longer and more complex the piece, the more work it’s likely to need to turn it into something you can publish. At a certain point, you might find it’s actually less cost-effective to generate a piece and have it professionally edited than to just have it produced by writers in the first place.

How we can work with generative AI

At Ascribe, we’re known for producing high-quality original content, but many of our clients also rely on us for editing services. We’ve taken long, complex technical reports and turned them into concise, accessible documents for a general audience. We’ve rewritten brochures, blog posts and other copy to incorporate a strong brand voice without altering the basic content. And we’ve made countless documents more engaging and scannable with basic line-by-line editing.

Whether we’re writing or editing, our end goal is the same: to produce highly appealing, readable content that delivers the right message in the right way for each client. And we can do the same for AI-generated content. We can do line-by-line editing for mechanics and fact checking, a stylistic update to ensure your brand is well represented, or a complete overhaul if the text isn’t quite what you had in mind.

To learn more about how we can help you put your best foot forward, contact us today.