Content AI – The Ups and Downs in Business


I am old enough to have left school around the time the internet emerged into widespread personal and business use. I started a website development business before the smartphone and was encouraging businesses to use social media (or Web 2.0 as we called it then) when MySpace was still the most popular. When I think back over my career about WOW moments that transformed how businesses can use technology, these have been them. We now have the next moment. Content AI, or more broadly AI in general as is applied to content generation, image generation, and the forthcoming music and video generation.

I am using the term Content AI in this article to encompass the many software that generates many forms of content based on text inputs. Examples of its potential use include blog posts, social media content, email content, position descriptions, book chapters, and fiction, the sky is the limit. I have been tinkering with an application called Jasper for almost a year and have more recently been able to compare it against ChatGPT.

You will have heard about ChatGPT, an application that uses artificial intelligence and natural language processing techniques to understand and generate human-like responses to text-based inputs. If you have had a “play” with it, which is what the first interaction feels like, you will have likely had your mind blown by its capabilities. You will also find yourself assessing its upsides and downsides.

With nearly twelve months of experimenting with it, I have a conflicted view. It has amazing benefits straight out of the box, and the potential is incredible. However, I am worried about the effect it will have on the professional development of developing marketing professionals, as well as anyone working in an industry where the ability to write creatively or technically is a skill to be honed and developed into one’s own strength. I might have had the same view of the pocket calculator when it was invented (it was before my time … just). I still have the same view about Canva.

Being a marketing agency, Content AI can dramatically reduce the amount of time it takes to create social media content and longer form content. However at what cost? The best analogy I can think of is letting a professional swimmer compete wearing flippers. They might get there faster but have they improved their skills and technique? I understand that times have changed and we make use of other AI software to reduce the amount of time and the application of personal expertise, but Content AI used in haste and for expediency, or to cover for individual shortcomings, poses the risk that budding content creators become at best pilots that drive the creation of content, or at worst, button pushers and copy and pasters. An AI Policy is being considered as part of employment conditions as one way to manage its use. I fear and accept that in the very near future, the ability to use ChatGPT will become an essential criterion in job advertisements for marketing professionals, especially for junior roles.

Enough about my own internal wranglings. Let’s look at Content AI itself. Whilst ChatGPT is now the most widely known and has brought Content AI into the mainstream, the technology has been around for a few years. I have been experimenting with Jasper for almost 12 months, with more recent comparisons and testing against ChatGPT. Jasper uses the same GPT-3 engine as ChatGPT, but the interface is structured toward content writers specifically. It has pre-built generators for the different types of marketing content and has inputs for tone of voice and SEO keywords. All of these are available for use in ChatGPT as well except you need to include these in your prompt. ChatGPT has many other uses (although Jasper now has a chat interface) including general information gathering, it can even write a happy country song including chord progression, and generate useful code for website development!

The downsides first

Content AI struggles with current events, and it warns that there may not be enough in the tank for anything occurring after 2021. This is sweet relief for journalists and content writers dealing with recent events. Secondly, because it effectively aggregates tone of voice it is hard for it to localise a tone of voice, e.g. how the world talks about wine is not how Hunter Region winemakers talk about wine. There are also inevitable factual errors and lapses in cultural sensitivities. A side note in regard to the accuracy, I asked Jasper to name the Australian team for the first test against England in 1977 and it got 7 from 11, ChatGPT got none correct.

There is the risk of duplicate content and SEO penalties. There are already rumblings that the Google Algorithm will be able to detect AI-generated content on websites and subsequently websites would be penalised for it. This is yet to play out or be tested, but if Google does implement this, it will come at the risk of human-generated content being penalised. It is a warning though to those generating website and blog content largely reliant on AI. In my opinion, ChatPGT poses the biggest risk Google has ever faced, which is why it has released its own version called Bard. Microsoft has invested heavily in ChatGPT integrating it now with Bing and its office products. It makes sense that Google will do what it can to ward off this competition. In regard to duplicate content, which can earn an SEO penalty, using AI-generated content could result in your content being largely identical to pre-existing content.

As I have written about previously, Google already penalises content written for search engines. However, if you can pilot your AI-generated content towards content that is helpful, it will not be an issue. Your content must demonstrate E-A-T (experience/expertise, authoritativeness, and trustworthiness). The danger with Content AI is that the outputs may deviate from that framework. Rigorous scrutiny and editing must be applied to it prior to publishing.

Risk of copyright infringement exists because AI generates content from a massive bank of existing content. There is a chance that a trademark or copyright could be infringed upon if large amounts of text are generated and unchanged from its primary source. There is also a reputation risk if AI content is used in your marketing content and contains factual errors and misleading information.

There is also a wider issue that concerns me a little, Because the engines use billions of documents produced the world over, how much of that was originally produced by the western world? And how much of that by white Anglo-Saxon males? What is the potential long-term impact on other cultural and minority voices and tones of voice?

What are the benefits?

I am simplifying them here as I know there will be plenty of opportunity to highlight how Content AI is going to benefit marketers and content creators in the future. I also save some demonstrations and examples for another day.

The biggest benefit is time-saving. These tools save A LOT of time. It probably takes me around 2 hours to research, write, edit, and publish a blog post. With the experiments I have undertaken so far, I can generate a reasonable blog post on a basic topic in around 15 to 20 minutes. This is time saved that can be devoted to other more strategic marketing activities and business development.

Another benefit is that it can help with writer’s block. Don’t know what to write about, ask the AI to come up with 10 blog post ideas for your particular service or product category. Once you choose a topic, you can ask it to start a post in a helpful tone of voice aimed at your target market and you are away.

Jasper in particular can create an entire campaign using the AIDA (attention, interest, desire, action) framework for content promoting a particular product or service that can be then rolled out over a campaign period. It isn’t great, but it saves time and beats writer’s block or idea stagnation.

A less positive benefit is that it lowers the barrier to entry for non-writers to be able to create content. Going back to my comment about seeing the ability to use ChatGPT becoming commonplace on job advertisements for marketing roles, expect to see it alongside ‘no marketing experience required’. However, it can offer small businesses without the resources to hire marketing agencies the ability to generate content to help with their marketing. I have offered encouragement to businesses to produce blog content for a long time, now add Content AI to the tool kit to help you to do so.

I have seen a post on LinkedIn from a Content and Social Media professional that you should think of Contact AI tools as your intern. I don’t mind that analogy, though the author has a self-interest. It can get things started and do a bunch of the grunt work and it learns and only gets better as it goes. The downside is that the operator doesn’t necessarily gain a lot of experience or skill development from it other than fine-tuning prompts.

How to manage it?

The danger with AI-generated content is that it can lack the uniqueness and the perspective and tone of the author and business. It is important to ensure that AI content has credibility and is of acceptable quality and follows Google’s helpful content guidelines. Perhaps it is a role of a marketing professional to apply quality control guidelines backed by brand guidelines including tone of voice. It must include humans in the content production process, and you can’t rely solely on AI to generate content.  Lastly, it goes without saying that fact-checking is essential.

I believe every business should put a policy in place regarding the use of AI for producing content and images on behalf of the company and clients. This policy would be in terms of having staff declare when content is AI generated for use and to include in the quality assurance process the checking of facts, accuracy, and corporate and regional tone of voice prior to publication.

I should point out that like all technologies before it, there are risks that come with efficiencies but there are also huge benefits. My fears of how it will impact certain professions are realities that have been faced and dealt with in the past. I should also point out that this post was written without the use of Content AI. You can probably tell as it sounds a bit like I am an old man yelling at The Cloud.

Image generated by Dall-E with the prompt “hyper-realistic person representing artificial intelligence”.


Disclaimer Note: You can take a look at Jasper via this link. It is a partner link, which I have never done before, however, this software is a game changer, so I am happy to promote it.


Latest from HYPERWEB

See what’s been happening here at HYPERWEB HQ and explore news and insights on web development, digital marketing, SEO and more.

View all +

AI in Marketing: The Human Touch Remains Key

It is not an understatement that AI tools have revolutionised the way marketers approach tasks such as data analytics, customer interactions, and content creation. However, as we embrace the efficiency…