Sorry bean counters: AI should bolster creatives, not replace them
After an investigation called out Sports Illustrated (SI) for publishing AI-generated articles under fake bylines, CEO Ross Levinsohn found himself out of a job.
There is some controversy over who was responsible for generating the stories. The Arena Group, SI's parent company, didn't specify a reason for Levinsohn's firing. In any case, people listed as the authors of the articles did not exist and the timing of Levinsohn's exit seemed more than coincidental.
Then there's content company Advon. In a detailed exposé by Futurism (a Knight Foundation grantee), reporters found publications across the Gannett and McClatchy properties used articles produced by Advon to fill their publications.
McClatchy used AI-generated articles attributed to at least 14 fake authors in more than 20 of its outlets, including the Miami Herald and Sacramento Bee, Futurism reported.
According to Futurism's findings, Advon originally used contract writers living in developing countries. In a bid to save more money, the company tasked those writers with critiquing the work generated by the AI. Eventually, the AI became sufficiently trained and Advon terminated its already underpaid authors.
After Futurism's report, McClatchy and Gannett deleted the generated articles and consigned the fake authors to digital purgatory.
As AI becomes more capable and mainstream, it should come as no shock that some executives are looking to it as a much lower-cost replacement for human creatives. While AI can replace some creative tasks, AI-generated content can also diminish quality and increase the generic-feeling nature of the resulting work. And that's to say nothing of the instances AI goes completely off the rails, like when Google's AI search recommended adding glue to pizza to make it stick better.
We'll examine the huge temptation AI offers when it comes to saving money, where and why AIs are limited when it comes to content generation, and legitimate ways AIs can help reduce time and cost without sacrificing originality and quality.
The temptations of employing AI
In my previous roles as software company president and magazine publisher, I hired a fair number of creative professionals. I had programmers, writers, editors, artists, and graphic designers on staff. I spent much of my time focused on one challenge: how to make payroll for the next week.
There's a wide range when it comes to how much creatives are compensated. Some writers get paid on piece work as contractors and barely make minimum wage. Others make six-figure incomes. The same is true for each of the creative professions. But even the least paid still often cost their company hundreds of dollars a week and thousands of dollars per month.
The managerial temptation is compelling. What if you could cut four out of five salaries (or the equivalent contract payments), and keep just one professional whose primary job it is to operate AI services (which, in total, often cost under $100/month)?
For most organizations, the people cost is the largest expense. Replacing humans with a couple of AI subscriptions might seem like the Holy Grail -- at least from the point of view of bean counters and others who focus on cost accounting.
As someone whose encore career is now thankfully out of the management arena and is squarely planted in the creative professional category, I can see both sides.
Also: I use Photoshop's AI tool every day - here are my 5 essential tips for the best results
For those of us who make our living as creative professionals, the prospect of being replaced by a bunch of blinking lights is terrifying. Some creatives know they produce enough unique value to justify their continuing employment. But those top-tier skills come with a bigger price tag for employers and clients, which creates yet another temptation paradox.
For outlets like ZDNET, for whom content quality is a competitive advantage, the cost of creatives is necessary and justified. For content farms, and those who hope to strike it rich through spamming affiliate links and crappy reviews all over the web, the need for human content creators seems far less so.
That brings us to a fundamental question: Do consumers of creative output prefer cheap or good? The answer is both. Some consumers gravitate to high-quality work. Others have little ability to distinguish between something of quality created by a human and something just churned out by an AI.
To be clear, not all creative output by humans is brushed with greatness. Humans phone in their work as much as the next AI. And not all creative output by AIs is low quality. Some of the images I created using Midjourney for my record album covers are just spectacular, far and away better than anything I could ever have done on my own, and I simply wasn't going to hire an artist for a personal side project.
While the promise of generative AI appears, on its surface, to be cheap content for everyone, there are costs, challenges, and limitations that prevent AI from being used to easily replace legions of knowledge workers.
AI's costs, challenges, and limitations
Here's something that seems like we're really living in the future: The surge of generative AI content to replace or impersonate what humans create is leading to the growth of another AI to combat that.
Yep, we live in a time of AI wars. In this case, we're talking about the ever-escalating attempts by low-quality volume content producers and spammers to fill the internet with money-generating garbage vs. Google, which has major efforts in place to lower the SEO value of automatically generated content.
Google's blog post on this makes for interesting reading, not only because SEO is so important, but because it showcases how creative work may be judged going into the future.
"Our focus is on the quality of content, rather than how content is produced," the post says. Google's algorithms are tuned to elevate original, high-quality content that demonstrates expertise, experience, authoritativeness, and trustworthiness.
While automation may have a role in, for example, posting up-to-date weather data or sports scores, detailed analysis benefits from a human voice. When it comes to the vast tracts of AI-generated filler designed to raise SEO value or trigger affiliate payments, Google actively fights this lower-quality content and spam.
Google's battle against poor-quality content alone should be enough to give most organizations pause. If you want to have a viable presence on Google, your content has to be good—and that alone helps protect the jobs of creative professionals.
There are other limits outside of concerns over Google SEO juice as well.
One of the biggest is that generative AIs just don't do big projects all that well. I've used ChatGPT to help me with my programming with great success. I've found it very helpful when I ask it for a small, very special-purpose routine. But whenever I've given it a more complex problem, it has failed.
The same is true for writing projects or even academic research. I would love to use ChatGPT to do a full literature review. But the best it usually does is cite one or two websites (and it loves Wikipedia).
Compare that to a real literature review done by a graduate student, digging through thousands of academic papers, journals, and other documents to create a carefully aggregated summary of available research.
There are also liability and legal costs as well. Here at ZDNET, we don't publish images produced by AIs because most of the image-generating AIs were trained on the entire internet, copyright be damned. We certainly don't want to be implicated or sued over the use of copyrighted content, just because it happens to be generated by DALL-E 3 or Midjourney.
The one exception to that rule is when we're reporting on how AIs work. In that context, we use AI-generated images for illustration purposes. Take this article I wrote last year about Midjourney vs. DALL-E 3. I was shocked to find that OpenAI's DALL-E 3 generated images of Snoopy. Yes, that Snoopy. Later in the article, it generated images that could have been pulled right out of The Nightmare Before Christmas.
Also: 8 ways to reduce ChatGPT hallucinations
Then there's the case of Missouri litigant Jonathan Karlen. He was fined $10,000 by an appeals court judge because he used an AI to generate dozens of non-existent citations in a legal brief filed in St. Charles County. Not only were some of his citations fake, but other aspects of his filing had errors, omissions, and formatting mistakes.
"The use of AI may bring about complex legal considerations including but not limited to intellectual property rights and compliance with regulations, which must be carefully navigated to avoid negative and potentially huge legal, financial, and reputational repercussions," said Carl D'Halluin, chief technology officer at data migration and protection company Datadobi.
We've talked about this many times: AIs make mistakes in the most authoritative-sounding ways. About a year ago, I spent an enjoyable if faintly evil afternoon convincing ChatGPT to go beyond its guardrails and lie, pants on fire and all. Don't try this at home, kids. I'm a professional.
The big takeaway here is that AIs make mistakes, as do humans. The use of AI-generated content could easily lead to both copyright claims and much more damaging liability claims.
But fear not. AI can be used by creatives as a force for good.
How creatives can use AI
Let's pause and think about some terms: skills, creativity, and vision. When I taught programming and multimedia (interactive animation and video) in college, most of my students were laser-focused on acquiring the skills that would land them specific jobs.
Skill is about the ability to do the task. Can you create a rough cut in Final Cut by stringing clips together? Can you write a JavaScript program to combine the name and URL of a web page into a bookmark on your browser? Can you create a simple graphic for a social media post?
All of these are about the ability to perform tasks with some level of expertise and precision. They're tasks necessary for doing the job, and in some gigs, they are the job. Humans have often created tools to help perform skills faster and with more precision.
For example, we use software-based video editors now, but back in the day, editors cut pieces of actual film together. Most woodworkers use power tools to get the job done faster, while some purists still prefer doing everything with hand tools.
When I got my first product manager job, we didn't have PowerPoint, but did a lot of presentations. We used old-school carousel projectors and 35mm slides. To put the graphics on those slides, we took weeks, and spent thousands of dollars per deck, working with outside slide production houses.
PowerPoint required those slide services to pivot. Not everyone survived. That's tech natural selection at work.
AI, at the skills level, is another power tool. It can get the job done faster, increase efficiency, reduce repetitiveness, and help lesser-skilled folks produce more skilled output.
Also: 1 out of 3 marketing teams have implemented AI in their workflows
But then there's creativity. Creativity is the ability to think of new ideas, new ways of doing things, and new projects. It's much more about imagination, the ability to think across disciplines, and the art of combining existing ideas in interesting and novel ways.
Vision is the ability to use creativity to see a path to a result. It's the ability to see and choose goals, articulate a mission, and describe desired outcomes.
Today, we can use AI to truly help with skills. Removing the background from a photo in Photoshop relies on AI, and it saves a ton of time compared to hand-specifying bezier curves to make that selection. AI can also help generate new ideas and even articulate vision. But it's not really capable of deciding what's good for you or your company.
"Creatives can leverage AI to handle repetitive data creation, customization, and management tasks, ultimately allowing them to focus more on art and innovation and less on administrative duties," Datadobi's D'Halluin told ZDNET.
The farther along you are in your career or your creative journey, the more time AIs can save you. This is because the more advanced you are, the more likely you'll be able to easily identify and define tasks that can be easily delegated, compared to tasks that require your own unique skills, perspective, and experience.
When it comes to corporate cost management, the real challenge lies in managers understanding their markets, unique offerings, and the competitive value their creative professionals provide.
Sure, AIs can reduce the workload of entry-level staffers and even take over some of the work that might have been delegated to those just entering the workforce.
While AIs can assist more senior professionals, years of experience and seasoning produce the key value. If you can't replace a senior creative with someone fresh out of college or art school, you can't replace that senior creative with an AI either.
AI and the future of creative work
An average of about 50 million Americans left their jobs in 2020 and 2021, according to a study from consultancy McKinsey & Company. Much of that was pandemic-era attrition, but there are still 10 million vacant jobs. As a quarter of Americans reach or exceed retirement age by the end of the decade, that number could grow.
While I'd never put it past a short-sighted bean counter to terminate productive workers in favor of a low-quality AI, that may not be the issue. AIs may help augment a workforce, allowing employees to keep up when their employers are unable to find more help.
McKinsey also estimates 30% of hours worked today could be automated by 2030. That could boost employee productivity, but it's also likely to disproportionately impact lower-wage workers, whose jobs are easier to automate or augment.
Meanwhile, a 2024 study by Microsoft and LinkedIn found that 78% of the knowledge workers surveyed are BYOAI (bring your own AI). In other words, they're not using corporate-sanctioned AI services, but logging into ChatGPT, Gemini, or Copilot. Nearly three out of four employees are using AI at work. Almost half of them started using it within the last six months.
Also: How does ChatGPT actually work?
In the survey, 66% of company leaders said they wouldn't hire someone without AI skills, and 71% said they'd hire a less experienced candidate who has AI chops over a more experienced candidate who is not AI savvy. Microsoft does have a horse in this race, but part of the reason it has invested so much in AI is because of metrics like these.
Personally, I have found AI to be a huge time saver. I initially started using ChatGPT and its ilk because I write about AI. But now that I have these tools in my kit, I'm never looking back.
While AI has failed miserably at writing full programs for me, there's no doubt ChatGPT has saved me weeks per project writing and assisted me in debugging small segments of code at a time. Midjourney has helped me create social media graphics for my wife's e-commerce business. ChatGPT's analytics processing has helped me do product market analysis I would never otherwise have had the time to do.
Fundamentally, I don't think companies will save money by replacing creatives with AI. However, I think companies would be remiss if they didn't recognize the performance and productivity benefits creatives can gain by augmenting their output using AI tools.
I'll leave you with an intriguing thought from Alex Ambroziak, senior creative producer of, content strategy at Shutterstock. He's coined the phrase the "generative AI paradox," wherein creatives who seem to have the most to lose to AI, may actually be the best qualified to use it.
"There's a misconception that AI, especially in its generative form, is simple to use and will get instant results but the reality is that the same skill set that is needed for composition and design is necessary to use AI creatively, only now, it's supercharged," he told ZDNET.
At least in 2024, AI is still a productivity tool to enhance and not replace.
You can follow my day-to-day project updates on social media. Be sure to subscribe to my weekly update newsletter, and follow me on Twitter/X at @DavidGewirtz, on Facebook at Facebook.com/DavidGewirtz, on Instagram at Instagram.com/DavidGewirtz, and on YouTube at YouTube.com/DavidGewirtzTV.