There’s no question marketing professionals are using artificial intelligence (AI) tools in lots of interesting ways, from creating social posts and short videos to generating blog post ideas and mining large data sets for insights.
There’s also no question that AI tools have advanced a lot in the past year.
But a big question on the minds of marketing professionals, from B2B startups to consumer brands to agencies, is: what does the future hold for AI in marketing?
Will it continue to help us work more efficiently and effectively—or will it make us superfluous and expendable?
Here’s a look at arguments from both perspectives, plus guidance for marketers on how they can maintain their relevance, based on what generative AI does, and does not, do well (per both humans and ChatGPT).
Content Authenticity Statement
Roughly 82% of this post was generated by me, the human. There are a few quotes and one longer section, duly noted, that were generated by ChatGPT.
Team Pessimism: AI Will Replace Marketing (and Marketers)
Heading up Team Pessimism is Christopher Penn, co-founder and chief data scientist at Trust Insights. In one of his recent newsletters, he referred to generative AI as an “extinction level event” for the marketing profession. He writes, in part:
Marketing as we know it is going extinct. Yeah, that’s a bold statement, but the reality is that generative AI has only begun to impact marketing, and we’re doing really dumb stuff with it…
Today’s AI models are capable of cranking out entire books in minutes. While many people use them at small scale, like writing an email or a blog post, today’s language models can generate entire working pieces of software and very, very long form content…
This is what a world of infinite content on demand looks like. Any content you want, machines will generate it for you when you want it, exactly how you want it, in the format you specify.
What’s notably absent from these scenarios? Us. Marketers, I mean. Other than a few ad spots here and there, there’s no opportunity for us to be participants in any of these AI-mediated interactions.
Marketing has been cut out of the picture in a world where generative AI can make you anything you want.”
He says a lot more, not all of it negative, so I’d encourage you to read his entire missive. Or watch the video version, if that’s more your thing.
Meanwhile, in 40 marketing and comms predictions for 2025, compiled by Frank Strong, Penn is joined by a couple of teammates:
Josh Inglis kicks it off for Team Pessimism, writing “You think to yourself – The PR AI agent is so productive I may not need to hire junior staff. And then you think even deeper – The PR AI agent is so productive, future clients may not need to hire me.” Ouch.
Bart Verhulst of Presspage takes it even further:
Over the last 6 months, I have noticed that many discussions on tech adoption centered on AI, and in particular, leadership’s wish to reduce costs. Read: cut the headcount of department…My prediction is therefore that we will see job cuts in comms departs and the need for PRs to showcase their value internally more than ever before.”
Team Optimism: Marketers (Who Embrace AI) Will be Fine
In the other corner, heading up Team Optimism (“Team Skepticism” might be more accurate) is Nobel Prize-winning MIT economist Daron Acemoglu. Spotlighted in Fast Company, Acemoglu projected in a research paper last year that “generative AI will only automate about 4.6% of tasks over the next decade.”
Acemoglu says that he’s not a pessimist, but is dismayed by what he views as excessive hype around AI. While he doesn’t address marketing specifically, he does shoot down a similarly dismal proposition about the future for lawyers:
They’re telling you, ‘Oh look at ChatGPT, it passed the bar exam, there’s going to be no need for lawyers,’ and all that crap, which has nothing to do with reality,’ he says. And that creates this environment, which I think is very, very bad, where CEOs and business leaders are feeling, ‘Oh, if I’m not investing in AI, I’m falling behind my competitors, I should just go ahead to find something to do with AI.’”
Frank Strong’s roundup of expert predictions actually featured a range of predictions on the impact AI will have on marketing and comms in the coming year, including some from influencers who seem to be on Team Optimism:
Michelle Garrett: “With the era of AI-based search upon us, PR matters more than ever because you want your business to show up EVERYWHERE. Why? The more your company is featured, the more likely it’ll be to appear when someone searches using a platform like Perplexity.”
Nicola Comelli: “Thanks to AI, a communications professional can simultaneously manage customized content for different platforms and audiences, while monitoring expands far beyond classic media to embrace a variety of sources. This is leading to a natural evolution towards a more comprehensive approach that intersects corporate affairs, reputation management and integrated communication.”
(Hmm…while neither makes the connection explicitly, these quotes seem to support the value of the WPO approach to marketing as AI use expands.)
Dr. Wendy Zajack (“Gen AI gets limited to a first draft”) and Carrie Eddins (“AI still can’t read the room”) also seem to be in the optimistic / skeptical corner for this debate.
How Marketers Can Avoid Death by AI
As the quotes above make clear, it’s hard to predict the future. But we can shape it. Knowing that B2B buyers will use AI as part of their decision-making process, perhaps marketers should actually help them do that, as Ardath Albee suggests.
Metaphorically speaking, the future isn’t a destination we’ve not yet arrived at, but more like a house we’ve not yet built. Can we reverse this trend (from another Chris Penn newsletter)?
Marketing (and marketers) will exist as long as we add value to organizations. As the capability of AI increases, marketers will need to figure out where it fits best, and where humans fit only. Here are two different aspects of the issue to think about.
Tasks AI Does Poorly, But Humans Do Well (According to Humans)
Based on both human and artificial intelligence, here’s a list of 14 tasks and types of content that AI doesn’t do well. These are, therefore, promising tasks for marketers to continue doing.
Original Research Reports
Humans and AI agree on this one. Generative AI, by its nature, can only regurgitate information it’s been trained on. It cannot produce new, unknown information. Original research naturally grabs our attention because we want to know what others think about a topic (think of the endless news coverage the latest polls leading up to the last presidential election).
AI can certainly help with specific aspects of the research process, but it takes human curiosity and judgment to create primary research studies.
Why AI struggles with original research, according to ChatGPT:
-
- AI cannot conduct primary research, such as designing experiments, collecting original data, or conducting fieldwork.
- It relies on pre-existing information and lacks the ability to independently generate novel data or validate hypotheses in the real world.
- Critical thinking and contextual understanding needed to interpret unique research findings are beyond AI’s current capabilities.
News Stories or Announcements
Another limitation that humans and AI agree on. In the media world, AI can’t produce on-the-ground journalism; that requires real humans. In the business world, AI can’t write an announcement introducing a new product or service, a key new hire, or a new sale or business partnership…because the model hasn’t been trained on any of that information. It is new to the world, and therefore requires human creation.
Why AI struggles with reporting the news, per ChatGPT:
-
- AI cannot provide real-time reporting from on-the-ground events, interviews, or firsthand observations.
- It relies on existing data, so its understanding of dynamic situations is limited.
Influencer Roundups
Consider the roundup of PR and comms predictions from Frank Strong referenced above. Or this roundup weighing in on the ideal number of experts for an influencer roundup post. Or this article from SmartMeetings sharing insights from some of the top women in event tech. Or any of the expert roundup posts linked here.
Again, AI may be useful in some parts of the process, but ultimately it takes a human to design and execute the process of collecting quotes, curating the responses, and of course real human experts / influencers to provide the content.
Metaphors / Compare-Contrast Pieces
Just as AI tools don’t too well with humor (see ChatGPT’s take below), it doesn’t do well with content that is clever or makes creative comparison, such as articles and blog posts on why digital marketing is like dating, chess, a rock band, or music. Or unusual associations like B2B marketing lessons from a trip to Disney World or learning X/Twitter strategy from your grandma’s adages.
Humans can make associations between unlike things quite naturally (and often humorously). Those comparisons are too illogical for AI.
Case Studies / Use Cases
Case study content is technically “produced” by the customer, but initiated, organized, and managed by marketing. Use cases come from a mix of real-world experiences and the knowledge of marketers and product developers.
In either case, it is “new to the world” content. AI can’t produce it because it hasn’t been trained on it. Again, AI could be used in the process (for example, to produce a rough transcript of a video interview) but it isn’t “creating” the content in any sense. Case studies, customer stories, and use cases can only be produced by humans.
Product Reviews (User-Generated Content)
Reviews, written by actual customers, are among the most impactful marketing content in many segments. While marketing professionals don’t directly create this content, they do enable and encourage customers to create it, and they (carefully) edit, then promote and repurpose that customer feedback once it’s created.
Of course, AI can be used to create fake “customer reviews”—but it can also be used to spot and remove them.
Tasks AI Does Poorly, But Humans Do Well (According to AI)
Per ChatGPT, “AI is highly capable of generating various types of content, but there are specific areas where its effectiveness is limited due to inherent constraints.” Here are eight more types of content that AI is generally not good at creating, according to AI tools, and the reasons why:
Highly Personal or Deeply Emotional Content
- Why AI Struggles:
- AI lacks genuine emotions, empathy, and personal experience, making it difficult to produce authentic, heartfelt content like personal essays, memoirs, or deeply reflective pieces.
- While it can mimic emotional language, the result may feel superficial or generic without the depth of human insight.
Abstract or Conceptual Art Descriptions
- Why AI Struggles:
- AI struggles with interpreting and articulating abstract concepts, such as symbolic art or philosophical ideas, especially when they require subjective or human-centered interpretations.
- It can describe literal elements but may fail to capture nuanced or metaphorical meanings.
Content Requiring Specialized Expertise
- Why AI Struggles:
- For complex or highly specialized domains (e.g., cutting-edge legal analysis, advanced medical diagnoses, or highly technical scientific theories), AI may lack access to the most current, specialized knowledge.
- It also cannot fully understand domain-specific nuances or implications without human oversight.
Culturally Nuanced or Sensitive Content
- Why AI Struggles:
- AI might unintentionally misrepresent or misinterpret cultural subtleties, humor, or norms, leading to content that feels tone-deaf or inappropriate.
- It lacks the lived experience needed to fully appreciate and navigate cultural intricacies.
Ethically or Morally Complex Arguments
- Why AI Struggles:
- AI cannot form moral judgments or engage in complex ethical reasoning that requires subjective values or human empathy.
- Its outputs might oversimplify or miss critical elements of a debate, leading to arguments that feel unbalanced or incomplete.
Novel Creative Work with Unique Perspectives
- Why AI Struggles:
- While AI can mimic creative writing styles or generate new combinations of ideas, it lacks the ability to develop original perspectives or groundbreaking ideas.
- True creativity often stems from lived experience, intuition, and a deep understanding of context, which AI lacks.
Highly-Regulated or Legal Documentation
- Why AI Struggles:
- AI-generated content may inadvertently omit critical legal or compliance details, leading to inaccuracies that could have serious consequences.
- Legal documents require precise interpretation of laws and case-specific nuances, which AI may not handle correctly without human validation.
Authentic Humor or Satire
- Why AI Struggles:
- Humor often depends on timing, cultural references, and an understanding of human absurdities, which AI struggles to grasp fully.
- Satire, in particular, relies on subtlety, irony, and the ability to critique societal norms—all challenging for AI.
Here’s why these limitations exist—again, per ChatGPT:
- Lack of Original Experience: AI generates content by analyzing existing patterns and data; it cannot create genuinely new ideas or insights from its own experience.
- Contextual Limitations: AI struggles with understanding deep context, subjective nuances, or abstract concepts that require a human touch.
- Dependency on Existing Data: AI is limited to the quality and scope of the data it is trained on or has access to, which may not include the most recent or specific information.
While AI is a powerful tool for many content-related tasks, these limitations highlight the ongoing need for human expertise, creativity, and judgment in areas where originality, emotional depth, or nuanced reasoning are critical.
What AI Does Well but Humans Need to Keep Producing Anyway – to Train AI
Unfortunately, as AI overviews answer more search queries, business websites will see an inevitable decline in organic search traffic. Fortunately, on the other hand, we (as marketers) have the power to impact the answers that AI search tools deliver—as long as we keep generating original content.
Here are seven types of content that AI is perfectly capable of producing, given a well-crafted prompt. But marketers need to continue creating all of these content types anyway, not only for promoting their products to human buyers but also to help train the AI models on what we want them to know.
“How To” Type Content, Guides, and Step-by-Step Instructions
Try searching Google for how to smoke a brisket. Take a look at the AI Overview at the top. I can tell you from smoking a lot of briskets (getting it wrong in seemingly every possible way before figuring out how to get it right), Google’s instructions aren’t quite perfect, but they are passable.
AI can provide a reasonably accurate step-by-step how-to guide for almost any procedure that doesn’t require highly specialized knowledge or skills (it won’t, for example, tell you how to build a nuclear bomb).
Competitive Product Comparisons
When Google dropped Universal Analytics and forced everyone to use the execrable GA4, I researched Google Analytics alternatives to use for clients. Many of the product sites featured head-to-head comparisons of their platform versus alternatives, for example:
Matomo vs Google Analytics (from Matomo)
Matomo vs Plausible (from Matomo)
Plausible vs Google Analytics (from Plausible)
Plausible vs Matomo (from Plausible)
Needless to say, the vendor creating the comparison matrix always comes out ahead in these comparisons. An AI tool or search engine (ChatGPT, Claude, Perplexity) could theoretically create a “vendor-neutral” comparison matrix. But where does it get its training data? From a variety of sources…including vendor websites.
Listicles and “Top” Lists
The web is full of articles and blog posts with titles like the Top 10 Marketing Tools to Step Up Your Game or The 22 Best Virtual Event Platforms. Again, these are relatively easy to create using AI. As marketers, we can’t control the AI output. But by creating and sharing our content far and wide, we can influence it.
Roundups of Statistics
These posts are very common as well, for example, 50+ Must-know social media marketing statistics for 2024 from Sprout Social. They are highly effective click bait and also tend to do well in search.
AI can be used to generate posts like this, but again it has to get the data from somewhere. If you can produce an original, authoritative post—or better yet, primary research—you’ll have a decent shot at showing up in AI Overviews.
Trends
People love posts with titles like “The Top Trends in (topic) for 2025.” And AI can definitely help spot trends.
The problem is, if the large language model (LLM) underpinning the AI is scouring the web, instead of a specific data set, it’s likely to come up with some nonsense—like gluing cheese to your pizza.
So, human judgment is definitely required. And human creation is still worthwhile, because AI is likely to rely as primary sources for its trends summary on…human-generated posts and articles about trends.
“What to Look For” Type Content
This type of content is applicable to any complex purchase, and very common in the B2B software world, for example, what to look for in an ERP system, or how to evaluate ERP software.
It’s common for vendors to write self-serving versions of these posts. Publishers like CIO, Forbes, and Tech Target will write (or allow vendors to write) their own versions.
AI could potentially scour all available sources to write up an “objective” list of attributes to look for, whether for ERP software or any complex purchase, but again, it has to get its information from somewhere…and that’s often from vendor websites (see “Competitive Product Comparisons” above).
Product Descriptions / Specifications
No one knows your product better than you, right? You are the authority. But again, AI could potentially come up with a more objective summation of your product’s features, functions, and capabilities, based on a broad array of sources.
True. But as the authority, you should be able to write a more thorough, comprehensive set of specifications than any other source. So, your output will be the primary training material for AI platforms. As a marketer, your primary source for this information will be your product engineers. But the final output is unquestionably marketing content.
Will AI Replace Marketing Professionals? Final Thoughts (Human and AI)
The experts who’ve pondered this question generally fall into one of two camps: Team Pessimism (Yes! Marketers are doomed!) or Team Optimism / Skepticism (No! AI is a useful and powerful technology, but it’s not magic—and it’s certainly not human).
There are types of content that humans produce well and AI does poorly, such as original research reports and influencer / expert roundups. There are other types of content, such as how-to guides and listicles, that AI that generate competently but human marketers need to keep generating anyway—in order to train the AI.
My crystal ball isn’t any better than yours, but my best guess is that what will be true for marketers is similar to what will be true for doctors, according to Jesse Ehrenfeld, president of the American Medical Association: “AI will never replace physicians — but physicians who use AI will replace those who don’t.”
Finally, what does ChatGPT have to say about this question? It’s a long answer, but to briefly summarize the “thoughts” of AI on this topic: “AI is unlikely to fully replace marketing professionals, but it will significantly transform their roles.”
AI is good at duties like automating routine tasks, personalizing content, generating insights from large data sets, and “generating content like…blog posts.” Yeah, be careful with that last one: AI can be a really smart assistant for writing professionals, but not a replacement.
According to ChatGPT, AI cannot replace creativity and innovation; strategic thinking; human relationships; ethical and cultural judgment; or contextual understanding (e.g., satire or sarcasm).
Ultimately, marketers who fear AI may be replaced—not by AI, but by marketing professionals who embrace and learn to creatively use it. I’ll close with ChatGPT’s final word on this subject: “Marketers who embrace AI and focus on their uniquely human strengths—creativity, strategy, empathy, and ethical judgment—will thrive in an AI-augmented industry.”
Hi Tom,
I think marketers that don’t want to learn how to use it effectively will be left behind. You make a great point about training the AI, it takes time to know you and what you want from the prompts.
AI has been amazing for me using these tools. I use at least 3 different ones and I’m working more than ever outputting a ton more work.
I edit and check stats and so much more before hitting that publish button.
Even on X I can edit posts and replies with Grok. It gives you ideas as your are posting for your own unique audience.
I can create videos from blog posts or make blog posts from videos.
However, it reminds me of when the computers first came out and we thought they would be no more paper. We all know how that turned out. ????
Great points Lisa. AI tools can do a lot to HELP marketers, but executives who believe the tools will REPLACE marketers will put their branding and lead gen at risk.
Hi Tom,
I agree that executives who believe the tools will REPLACE marketers will put their branding and lead gen at risk. Good and abundant human content creation helps create the information that AI is sourcing and disbursing.
Scott Flynn
Best Corporate Events
Absolutely, Scott. AI keeps getting smarter, but it will never be original or authentic, and it relies on new information — created by humans — for its never-ending training.