The True Cost of AI Generated Content Marketing

A tectonic shift is taking place in content marketing. More firms than ever are adopting AI tools. According to CMO Survey in a study from Spring 2024, 60.4% of respondents confirmed that their business had been using artificial intelligence for marketing for less than one year. Naturally, this means that many businesses are making new investments in the technology. Whether you’re already in the ring or are considering adopting these tools for your team, you need to understand the true cost of AI generated content marketing. What follows are my findings as a professional content marketer to answer these questions: how does AI generated content affect your search ranking, how does it affect your content quality, and what’s the most cost-effective way to incorporate AI into your content strategy?

Risk Assessment: the AI Wager

Let’s start with some assumptions. The questions you’re looking to answer are risk assessment question. In my opinion, there is no better risk assessment question than Pascal’s Wager — you either have eternal suffering or eternal happiness, or you could gain or lose nothing significant. Pascal’s Wager is all about how we know nothing about whether a god exists and identifying the most beneficial path to take to maximize our chances of happiness.

Although we can use a number of resources to understand how AI generated content can hurt your business, there are a lot of factors that are out of our control, such as:

  • Possible SERP Penalties: if search engines detect AI content, they could decide to penalize your pages.
  • Misinformation: AI generated content can be incorrect, resulting in misinformation and low-quality content that may be less engaging or harmful to your audience.
  • Originality and Quality: if you deliver valuable content to your audience, you will have a higher conversion rate. Low quality, AI generated content can sabotage this objective.

So, assuming that we know nothing about the potential impacts of AI content marketing, let’s create a quick matrix to visualize the possible outcomes:

Similar to Pascal’s Wager, there are two benign outcomes, and two ‘extreme’ outcomes depending on what we choose. If we don’t use AI, not a lot changes. This could be good or bad, depending on how you want to grow your business.

Assessing the Outcomes

However, things get serious when you consider whether AI-generated content will be penalized or will strongly affect your content marketing strategy in some way. Let’s say that search engines do penalize content made with AI. If you use it, your content could be severely disadvantaged compared to competitors. However, if they aren’t penalizing it and it doesn’t affect your content’s value, you could have a substantial opportunity.

Therefore, if you are looking for a conservative, less risky approach to content marketing, the best outcome according to these assumptions is to not use AI. However, markets have become far more competitive during AI’s hype cycle. If you want your business to continue growing, you may need to take some risk to remain relevant.

Some of these risks can be mitigated. However, to understand what that means for your business, we need to examine each risk a bit further. Let’s take a moment to investigate a few of the problems I noted earlier to see how we might approach them when utilizing AI in content marketing.

Search Engine Penalties

Let’s focus on Google: does Google penalize AI-generated content? The answer isn’t so clear cut. Interestingly, Google doesn’t say anything about AI-generated content in its recent announcement about March 2024 indexing updates. However, it does affirm penalties to low quality content, including ‘auto generated content’. Particularly, I think we should focus on this specific part of Google’s announcement:

We’ve long had a policy against using automation to generate low-quality or unoriginal content at scale with the goal of manipulating search rankings. This policy was originally designed to address instances of content being generated at scale where it was clear that automation was involved.

Today, scaled content creation methods are more sophisticated, and whether content is created purely through automation isn’t always as clear. To better address these techniques, we’re strengthening our policy to focus on this abusive behavior — producing content at scale to boost search ranking — whether automation, humans or a combination are involved. This will allow us to take action on more types of content with little to no value created at scale, like pages that pretend to have answers to popular searches but fail to deliver helpful content.

From Google’s March 2024 Announcement

This supports earlier Google doctrines like their February 2023 post specifically explaining that AI-generated content wouldn’t be blacklisted in Google search by default. In that post, Google asserts that automatically generated content can have great value to end users. With this in mind, consider these two scenarios:

Neither of these are bad, but let’s determine which one will perform better in Google search based on the language Google used in their announcement. According to the announcement, Google is penalizing content with “little to no value created at scale”. Therefore, we should think about the value of the content to the end user and the originality of the content.

Scenario 1: Automatically Generated Pages

In Scenario 1, there is a possibility that the content could be valuable to the end user. With there being so many choices for consumers, comparing cell phone specifications is something that can quickly overwhelm shoppers. Using AI has obvious advantages at scale for collecting, summarizing, and displaying this content on a website. However, I believe that the use case I described could have significant risk of being penalized by Google’s algorithm. Since the data is coming from other sources, it’s possible that Google’s algorithm may not determine the content as being original or useful.

UpPhone.com is a good example of automatically generated pages at scale. However, they don’t use AI to write each page. Their original value proposition is their comparison tool that helps consumers choose the cell phone that’s best for them.

That being said, the business in Scenario 1 could find ways to make the content more original. For example, maybe you supplement each page with human reviews, images, and comparisons. This could help make the content less homogenous across the site and offer more value. However, arguing with Google’s bots will be an uphill battle since you are still generating your content automatically at scale. This could be the case even if you don’t use generative AI at all.

Scenario 2: Supplementing Content Strategy with AI

Scenario 2 is a bit different. This example doesn’t involve scalable, automatically generated pages. Instead, a marketing manager approaches their audience’s pain point more directly. Let’s say they’ve identified five key problems that customers are having with their SaaS product. They then build ordered lists of step-by-step solutions from discussions with software engineers. Then, they can give a generative AI tool like ChatGPT or Microsoft Copilot those same steps to create either an entire article or a draft.

How you use AI in this scenario will directly affect the results. However, I believe that Google will be much less likely to penalize this kind of content compared to an approach that uses automatically generated webpages at scale.

As I said, neither of these scenarios are bad. Each has their own unique challenges. However, with regards to Google search, Scenario 1 faces more of a challenge than the latter example simply due to Google’s greater attention placed on automatically generated pages at scale. According to Google’s language, generative AI will not disqualify a page from Google search by default. So, with that in mind, it’s time to move on to the other risks and costs of AI in content marketing.

Content Quality & Originality

Google lists an FAQ question on their February 2023 post that I linked to earlier that asks if AI-generated content can rank more highly than other content. They explain that “Using AI doesn’t give content any special gains. It’s just content.” However, AI generated content has one fatal flaw: its quality.

According to a study conducted by Sun Y. et al (2024), large language models have a tendency to produce incorrect information we can call hallucinations. Over a series of 284 tests, the researchers identified many types of errors and categorized them into eight types. Below is a summary of the largest types of error by category:

Most common AI generation errors sun y et al. 2024 Mathematical errors, Measurement unit errors, 40, Reasoning errors, Physical reasoning errors, 39, Factual errors, Objective fact errors, 24, Logic errors, Contradictions, 22, Overfitting, Falling into traps, 11, Text output errors, Spelling errors, 8, Other errors, Restrictive filtering, 6, Unfounded fabrication, False academic information, 4

Overreliance on generative AI processes for content generation will result in these errors. Your goal is to produce valuable, high-quality content for your audience and customers. To do that, you need to account for these errors somehow. Otherwise, you won’t have just search engine penalties to worry about — people who do find your page will just click away. Additionally, when you publish errors created by AI, you contribute to misinformation, which can have serious consequences on our society.

Also consider plagiarism. Google prioritizes original content when indexing. It’s no secret that businesses train large language models on large swathes of copyrighted materials. Utilizing generative AI for content marketing comes with a risk of unintentional plagiarism which could hurt your search performance and cause ethical and copyright concerns for your business.

Choose Your Approach

Now that we’ve done a bit more digging into the various risks that AI generated content could pose for your business, it’s time to choose your approach for content marketing. Generally speaking, there are three approaches to creating content, each with their own level of risk: no AI, some AI, and all AI.

A sliding scale from less risk to greater risk. The least risk is no AI. In between is some AI. The most risk is all AI.

Choosing not to use AI naturally has the least risk. Despite being slower, it’s much easier to ensure that it’s original. On the other hand, creating content with AI is the most scalable and fastest solution. AI is much more efficient at writing content than humans, but it also has the greatest degree of risk. Depending on the nature of your content, AI puts you at risk of penalties from search engines and inhibits the quality of your work.

Treading a road somewhere in between these two extremes may be the best move for many businesses. You don’t need to use AI for every part of your content marketing process. You could use it for idea generation, outline creation, drafting, and even creating pieces or the entirety of certain articles. However, it’s critical that a human oversees the entire process and moves through each step one at a time. Think of it like being in the seat of a self-driving car. The car will make mistakes and doesn’t know where it needs to go. It’s up to you, the marketer, to drive the vehicle of your AI content tools to your destination. If you can do that, you can still produce valuable content that can rank on search engines.

Mitigating Penalization

Now, I’ve already outlined what Google has said about AI-generated content. However, despite what I’ve read, I still have a bad feeling about this. Google has not been entirely honest with us in the past about how their search engine indexes content. The ongoing deliberations on how Google will be punished for its monopoly on search are a product of this dishonesty. Given Google’s obsession with data, I wouldn’t be surprised if they are using AI detection tools to understand just how much of the web is generated by AI. Then, either now or at some point in the future, they may decide to prioritize human-created content.

Because of this, I highly recommend that if you are using any AI tools in your content marketing processes that you work to humanize it as much as possible in an honest way. Although there are AI tools out there that modify your content to remove LLM watermarks and try to make it undetectable by AI detection tools, I don’t think this is the best approach. You need to make sure that the quality of the material is as high as possible, and that can only be done by the same kinds of people that are in your audience: humans.

Use AI Detection Tools for Auditing

I highly recommend having all your AI generated work thrown through an AI detection tool like ZeroGPT. Get a percentage output from the tool and write it down. Then, you or another human writer should go through the article and rewrite it (or at least parts of it). Then, throw it back through the tool to see the new result. You should aim to reduce it by a certain benchmark. Seeing as AI detection tools can sometimes have false positives, the kind of benchmark you’re looking for may vary on your use case.

A screenshot of the nogpt website with the previous two paragraphs of the article you just read. The tool falsely believes that 93% of the text was AI generated, with 7% human written.
Hilariously proving my point, another tool called NoGPT thinks the last two paragraphs were generated by AI when I came up with all of that by hand.

This is a great opportunity as well for humans to quickly check AI outputs for mistakes, hallucinations, and quality issues.

Humans Value Human Content

I have one last thing to point out before we wrap up related to the value of your content. In all the studies I’ve seen so far, humans greatly value human content over AI-generated content, and naturally distrust AI-generated content.

Survey of academic and creative communities on challenges AI poses (Masoud neyani et al. 2024) Aspect of Challenge, Strongly Challenges, Somewhat Challenges, Neutral, Not a Challenge, No Opinion, Challenging Authorship, 20%, 45%, 20%, 10%, 5%, Challenging Creativity, 15%, 40%, 25%, 15%, 5%, Intellectual Property Rights, 25%, 35%, 20%, 15%, 5%, Addressing Ethical Concerns, 15%, 45%, 20%, 15%, 5%, Addressing Legal Concerns, 10%, 35%, 30%, 20%, 5%
Results of a survey by Masoud Neyani et al. (2024) on academic audiences’ opinions of how AI challenges various topics.

AI is an immensely powerful tool, but you need to remember the purpose of content marketing: to share valuable and relevant content to your audience that will help build a better relationship between them and your product or service. Your audience are humans. AI may be able to help you make that connection, but it cannot make that connection for you. There is no denying that content marketing is a social strategy that requires conversation, connection, and human oversight. It’s not as simple as automating your content like a factory. Keep this in mind when you choose your next campaign strategy.

Keep Your Content Human

Good news! I’m a human, and I’ve been helping teams like yours connect with audiences for years. Whether you’re using AI and need a bit of the human touch, or if you need a content strategy developed from scratch — I’m here to help. I’ll help you reach your benchmark for AI detection tools — as low as we can go. I’ll tell your brand’s story, help you rank in search engines, and make blog posts and landing pages that convert. All you have to do to take that first step on your journey is to reach out to chat!

Liam Shotwell: SaaS Content Writer

I’m a human B2B SaaS content writer with a focus in SEO. I’ll shine a light on your brand to help you find potential leads lost in the dark of SERP.

See My Portfolio
Reach Out

Attribution

Icons in this piece created by Maxim Basinski Premium, Freepik, and Nurlaili.