How the Content Marketing Landscape Has Changed With AI 

In late 2022 and early 2023, the launch of tools like ChatGPT, Google Bard, Perplexity AI, and Claude sparked an explosion of interest in AI. They seemed poised to revolutionize the business landscape, especially in the creative, marketing, and tech industries.

Hailed as game-changers by some and the beginning of the end by others, these tools seemed ready to streamline content creation. Would it sideline traditional methods of content generation? How would Google react to AI-generated content? Would SEO be reshaped and redefined?

Generative AI tools ignited a wave of excitement among marketing professionals, sparking speculation and experimentation. personal loans. Representing a new evolution of technology with the potential to generate content at unprecedented speeds, many marketers and knowledge professionals explored their capabilities.

Those who engaged with AI tools discovered that, while they are impressive, they are not infallible. They are prone to inaccuracies, inconsistencies, and hallucinations. Additionally, since the AI writers are rehashing pre-existing information, attempting to create fresh, authoritative content from these tools was challenging.

Despite these hurdles, AI’s potential is undeniable. The insights gained in 2023 have set the stage for an exciting, evolving journey. In this white paper, we dive into these discoveries, shedding light on how generative AI is impacting content creation and SEO, including how marketers are harnessing AI’s potential while navigating its limitations.

Do you want to share this story? Click to download it as a PDF.

The Burst of Generative AI

Generative AI (gen AI) captivated public attention at the beginning of 2023. OpenAI’s ChatGPT achieved one million active users within two months of its launch on November 30, 2022. After earning gross revenues of less than $28 million in 2022, OpenAI is generating $80 million a month, on target for $1 billion in 2024.

While it’s difficult to estimate the precise number of AI writing tools that emerged in 2023 due to the rapid acceleration of the industry, the global AI market grew from $95 million in 2021 to over 207 million in 2023. In February, searches of the term “generative AI” shot past a previous high of 34 index points in December 2022 to 100.

Google: Quality Over Source, AI Content Not Penalized

Marketing teams started the year cautiously testing gen AI’s capabilities, waiting to hear how Google would treat AI-generated content. On February 8, 2023, Google announced it would prioritize the quality of content over how it’s produced. Content creators felt they had a green light to test the new technology.

Marketers Dive into AI Experimentation

Depending on their vertical, marketers found ways to implement the new applications. AI chatbots could reply faster to customer queries than human agents. Tools like Claude could summarize long articles or video transcripts in the blink of an eye. ChatGPT was an invaluable brainstorming partner, particularly in content strategy development or tasks requiring a rapid ideation process.

A 2023 survey of 900 marketers revealed that:

• 77% of respondents want AI to reduce time on repetitive tasks
• 98% of marketers use AI tools for some tasks
• 45% said they are experimenting with the tools
• 29% say the tools are infused into their daily workflows

Not “Plug and Play”

Experienced marketers soon realized the new AI writing tools were not “plug and play.” The quality of the output relied on the skill and knowledge of the individual crafting the prompts.

As described in a Tweet by OpenAI’s CEO, Sam Altman, around the time of ChatGPT’s December 2022 launch, “ChatGPT is incredibly limited, but good enough at some things to create a misleading impression of greatness. It’s a mistake to be relying on it for anything important right now. It’s a preview of progress; we have lots of work to do on robustness and truthfulness.”

Discovering AI’s Limits in Content Creation

As marketers delved into the world of generative AI tools, their limitations became apparent. This was especially true in content creation. From nuanced inaccuracies to challenges in originality, the initial excitement gave way to a deeper understanding of these tools’ capabilities and constraints. 

Fabricated Content

Unreliable AI-generated “news” websites with little human oversight have propagated this past year. NewsGuard’s team identified 557 poorly monitored AI “news” sites spanning 15 languages, from repackaging mainstream news without citing sources to spreading misinformation, such as an entirely fabricated TikTok video of Paris streets littered with trash that got more than 400,000 views in a month.

Confidentiality

Confidential information used in a ChatGPT prompt can be leaked. While an individual account may seem private, Samsung reported three instances where employees unwittingly leaked sensitive information by inputting it into ChatGPT. Microsoft employees have been restricted from using ChatGPT due to security concerns despite Microsoft’s considerable investment in OpenAI and OpenAI’s assurance that security measures have been put in place.

Copyright Protection

Generative AI is essentially a very advanced auto-complete tool that uses information already published on the web as its source. It lacks the context, social awareness, and emotional intelligence of human writers. This brings up a few complex questions that are not yet fully resolved. For example:

1. Who owns AI-generated output?
2. How do you guarantee that AI content isn’t plagiarized?
3. What data set or large language model (LLM) is being used to produce this content?

As of August 18, 2023, a ruling by Judge Beryl A. Howell of the DC District Court ruled, “United States copyright law protects only works of human creation. Human authorship is a bedrock requirement of copyright.” This meant that content produced by an AI tool without human involvement would not be eligible for copyright protection.

Lack of Originality

Content generated by an AI tool is not technically plagiarism and will not usually be flagged by a plagiarism detection tool because it doesn’t copy exact sentences and phrases. However, it is paraphrasing existing content.

RevolutionParts, a fast-growing tech company supporting automotive dealers, spent much of 2023 experimenting with generative AI tools. Thomas Tourville, Manager of Digital Marketing, said, “We tested AI content creation on over 30 websites, and every piece of content has been detected as plagiarized or AI content. Additionally, when vetting the content’s accuracy about a particular subject, 73% of the total pieces of content came back as inaccurate.”

The risk of plagiarism in AI-generated content, then, occurs at a larger scale. If your content is a rehashing of what is already written, is it the unique, authoritative content that Google is committed to rewarding?

Well, Revolution Parts may have an answer to that question. Tourville goes on to say, “This study was conducted over a 3-month period in mid-2023, across 500+ pages of content, including blogs, category pages, and product pages. None of these sites or pages have experienced any positive lift in performance or surfaceability on SERPs as of December 2023.”

Google’s February response to AI-generated content explains that it has been dealing with mass-produced, low-quality content for years, and the advent of new AI tools doesn’t change the fundamentals of its approach. The guidelines state, “However content is produced, those seeking success in Google Search should be looking to produce original, high-quality, people-first content demonstrating qualities of E-E-A-T.”

The Value of Intelligent Oversight

The power of a tool lies in the hands of its user. Consider a scalpel: it’s a life-saving instrument in the skilled hands of a surgeon. Without expertise, it’s merely a piece of metal. This principle holds true with AI-driven content creation.

As MarketingProfs founder Ann Hadley wrote, “You can’t put AI in the hands of someone who can’t write and expect anything other than mediocre. AI doesn’t work on its own… not really. It requires discerning people who care.” 

AI tools are not a standalone solution. They are potentially time-saving tools for human intelligence and creativity, a critical understanding for marketers looking at the impact of Google’s E-E-A-T and YMYL frameworks.

E-E-A-T Framework

The E-E-A-T (Expertise, Authoritativeness, Trustworthiness, and Experience) framework continues to drive search rankings. Google and other search engines are committed to delivering search results that accurately and completely answer people’s questions from a position of authority and expertise. To do this, they prioritize high-value, unique content written by knowledgeable people.

The question of AI-generated content also ties into Google’s August 18, 2022, Helpful Content Update. It was designed to reward people-first content that searchers find useful. Google’s intent included reducing traffic to pages built to drive clicks without delivering value.

Essential to Google’s criteria of helpful content is original information that provides substantial value. Content that demonstrates expertise, establishes authority, and earns trust is not just preferred; it’s rewarded.

YMYL Content

Google defines YMYL (Your Money Your Life) businesses as those related to people’s money, health, or overall livelihood. For example, financial planning and healthcare. Due to the impact these businesses can have on people’s lives, Google holds YMYL content to a higher standard.

Emily O’Connor Kefs, our CEO at ContentWriters, explains that, “Your Money or Your Life” (YMYL) topics require a deep understanding of complex concepts, and associated content should be backed by reliable sources and analysis.”

Kristina Ackermann, Director of Content and SEO at American Addiction Centers, expands on this point, saying, “There simply is no substitute for an educated professional with a background in the field they’re writing about… and we’ve found through our experiments with generative AI that these tools lack the ability to distinguish medical journals from the fray of the internet at large.”

AI’s limits stand out in this category because gen AI delivers output based on its input. For ChatGPT, that was 300 billion words. While that is an astonishing database, it is only rehashing existing information and will not add to the unique content on the web.

The USC Libraries research guide outlines the known limits of gen AI, which include:

• Gen AI models are not databases of knowledge. They are a technology that attempts to synthesize and reproduce information it was trained on.
• Many gen AI models like ChatGPT are trained on data up to a specific date, which can be a couple of years in the past. This results in outdated information or gaps in knowledge about current information and events.
• Validating and properly attributing the source of the content from gen AI is difficult to impossible.
• AI tools may not be able to update or verify the content they generate.
• Depending on the prompts given, the output of gen AI models can be oversimplified, reductive, or generic.
• Large language models (LLMs) can generate fictitious or incorrect information and present it as factual or accurate.

On the subject of using gen AI for content creation, Rand Fishkin, Cofounder of SparkToro and Snackbar Studio commented, “It makes sense to me that AI tools in marketing (especially content ideation & creation) saw a huge spike, then a drop and leveling-off. It can be “fine,” but it’s never incredible.”

It’s not only about the quality of sources when it comes to producing YMYL content. It’s also about having reputable writers connect on a human level with your audience. Ackerman stated, “Working with human authors gives us an advantage in providing readers with information they can trust. We’ve seen this demonstrated not only in our search rankings, but in our patients and their loved ones as they work with our staff to determine their care.”

AI Detection Software Emerges

With the explosion of ChatGPT and other generative AI tools, several industries, from journalism to education, found themselves in an awkward position. Was the content they were receiving written by humans or generated by AI?

Policing AI Usage

Scientific journals rushed to preserve the integrity of their publications under concerns that the quality and reliability of their content were threatened. Some journals banned content generated by AI writing tools, while others limited how it could be used and required disclosure about which tools were used, how, and why.

Journalists experimented with ChatGPT and other generative AI writing tools, determining their usefulness for their publications. Wired’s AI disclosure states that after working with the tools, they discovered that AI writing lacks the quality and accuracy they use in their reporting, but they will use it for marketing emails and brainstorming headlines, social posts, or story ideas.

Teachers were suddenly faced with a tool that could write students assignments for them, making assessing a student’s skills and comprehension much harder. New York City banned access to generative AI by students and teachers, saying they could help children cheat. To compensate for the downside of AI tools in education, some educators are proposing a “flipped classroom” where learning is done outside of class and practice is done under supervision.

In the business world, companies that rely on content marketing either set a policy disallowing writers to use generative AI or requiring its disclosure. However, to uphold those policies, they needed a reliable and robust AI detection tool.

What Distinguishes AI From Human Writing?

Aside from inaccuracy, the most common issues with AI-generated content are a lack of depth and personality and repetitive language. Human touch impacts content quality, especially for businesses whose goals include building the trust and authenticity essential to search engine rankings. Organizations needed a way to verify their content was reliable, accurate, and original.

A Princeton student named Edward Tian identified two factors that more reliably define human writing. They are “burstiness” and “perplexity.” Perplexity is a measure of randomness within a sentence, and burstiness is the overall randomness of sentence structure in the text. Humans tend to have more variation in their sentences than gen AI tools.

Challenges to AI Detection

The first wave of AI detection tools couldn’t provide the reliability that organizations needed. ChatGPT shut down its detection tool after six months, citing a low accuracy rate. Other tools unfairly penalized non-native speakers or were easily fooled.

With the range of human writing styles, variations in writing and language skills, and different levels of subject-matter expertise, constructing a reliable AI detection tool challenged programmers. A tool’s ability to detect was also impacted by the user’s prompts. For example, including “write in the style of the New York Times” might produce a more human-like output than simply saying, “write a blog about growing culinary snails.”

Originality.ai

Originality.ai emerged as one of the most reliably accurate AI writing detectors. Originality.ai checks for the uniqueness of text against public-facing web content combined with an AI checker.

However, even the best AI detection software can give a false positive. In academia, this can result in students being accused of cheating when they did not. Even though they aren’t 100% perfect, AI detection tools vastly improve an organization’s ability to uphold its AI policies confidently.

If a writer tries to use AI to paraphrase previously published human-written content, Originality.ai can detect it. Content that is written by AI and edited by a person may still be flagged by AI detection software, depending on the amount of editing and rewriting done by the human writer.

Ultimately, Originality.ai serves as a tool to inform and protect the audience and the publishers. Madeleine Lambert, Director of Marketing and Sales at Originality.ai told CW, that their company “believe[s] in the ethical and transparent use of generative AI…generative AI has its place, but [it] should never be passed off as human and/or published unbeknownst to the owner of that content.”

This is a fair approach that helps readers navigate an ever-saturated world of content and protects editors and marketers who publish a high volume of content from different places. 

Finding a Happy Balance

What is the role of gen AI in content strategy, and how can it be used to keep up with marketing trends? Although gen AI is still a developing technology, it holds the potential to streamline workflows and speed up repetitive tasks.

Organizations Articulate Their Standpoint

As with all technological advances, some organizations and professionals avoid it altogether while others use it enthusiastically. Institutions like Harvard, the city of San Jose, and the publisher Elsevier published clear AI policies outlining acceptable use and the responsibilities of the author.

All three warn users to protect confidential information and affirm that authors are responsible for the accuracy and uniqueness of their content. Elsevier expressly states AI cannot be cited as a co-author. It prohibits the use of gen AI technologies to create book covers or graphics and instructs authors to disclose whether and how AI writing tools were used in their process.

Some Marketing Companies Find Ways to Leverage AI

Other marketing companies have embraced AI technologies to expand their service capacity. Unbounce, a landing page software company, has integrated AI to help clients A/B test their landing pages more rapidly. Instead of testing one feature at a time, the AI tool can help speed up the optimization process.

Kaitlyn Bose, co-founder of the SEO agency Frac.tl has taken the process a step further, writing and generating code to guide ChatGPT to execute tasks that her team does daily. Her publicly shared code performs tasks like identifying keyword gaps, monitoring news stories, and optimizing onsite links.

In content marketing, professionals are still discovering how to use gen AI best. HubSpot’s survey of 1,350 professionals revealed the top five ways marketers are using AI:

1. Generate ideas (33%)
2. Write copy (28%)
3. Create marketing images (26%)
4. Summarize text into key points (25%)
5. Translate text into different languages (23%)

An Ongoing Endeavor

Can generative AI produce original content that provides value? The answer to that question depends on the proficiency of the user, their clarity on the subject, the amount of information already on the web on that subject, and the skillfulness of the prompts.

Derek Johnson, ContentWriter’s Director of Content Strategy, explains, “The struggle is with quality control and project management. I have these conversations all the time now with businesses asking us to edit high volumes of AI-generated content.”

Writers may use AI tools to help with headlines, topic lists, outlines, writing, or editing. Some use the tools to create drafts, striving to add the human touch and align the text with brand guidelines while revising. The most essential step for any content creators using AI is to fact-check their content.

Looking Forward to 2024

If 2023 demonstrated anything, it was that getting content on the internet is easy and can be done cost-effectively. This ease makes content that speaks to the audience, informs, and inspires even more valuable.

Rather than replacing quality writers, AI tools are accelerating the need for an effective writing strategy. What’s needed are writers with a deeper understanding of a business’s audience, a commitment to authenticity, and a dedication to quality.

Human Creativity is Irreplaceable

When Princeton student Tian was building his AI detector, he repeatedly scanned a New Yorker article by author John McPhee titled “Frame of Reference.” McPhee was a pioneer of creative nonfiction, known for making complex topics accessible and informative. Through all of Tian’s testing, McPhee’s article consistently came up as the “most human” on all metrics.

When reporters asked McPhee why that might be, he said his writing went beyond reporting facts or explaining science. He saw through the eyes of the expert and wrote from there.

People infuse writing with human emotion, a unique voice, in-depth research, and personal experience. Kefs says, “We’re excited about the possibilities generated by AI content marketing, but we’re also aware that it’s not a silver bullet. One thing remains clear: The value of human insight, creativity, and expertise is irreplaceable.”

Quality Will Remain Paramount

Search engines will continue to reward quality content that provides reliable information. John Mu, previously a Google Senior Webmaster Search Trends Analyst, recommends posting an author’s name whenever possible. When anonymous, generic content is almost effortless to create, bylined articles and credentialed writers with verifiable expertise and experience become more valuable.

Moving into 2024 and beyond, the content that will stand out is well-researched, thoughtfully written, and genuinely informative. Companies that recognize and adhere to these values will thrive in search rankings and build lasting relationships with their audiences.

Not a Fleeting Trend

AI is not a fleeting trend, it represents a significant technological evolution. The conversation about its best use and implications is just beginning.

From personalizing content, streamlining and tailoring customer communications, analyzing data and trends, and identifying marketing opportunities, gen AI offers exciting possibilities for marketing and content creation. Even so, it is a tool and a technology, not a replacement. The heart of compelling content remains the insights, experiences, and expertise that only people can provide.

ContentWriters’ Approach to AI

As a writing service dedicated to delivering original, authoritative, and SEO-optimized content, ContentWriters incorporates Originality.ai into its workflow. Every piece is scanned; our expert writers know we do not tolerate AI-generated content. 

For our clients, we provide editing services for AI-generated drafts, so that the final output is accurate and meets their brand guidelines. We are dedicated to delivering content that stands out for its quality and human expertise.

Joan Weisman

Joan Weisman’s strong research skills, SEO knowledge, adaptability of style and voice, and passion for staying current make her an asset for freelance web content, SEO, and copywriting.