What has been said about artificial intelligence, better known as AI?
• “AI has the potential to be more transformative than electricity or fire.” — Sundar Pichai, chief executive officer, Google
• “AI will not replace humans, but those who use AI will replace those who don’t.” — Ginni Rometty, former CEO, IBM
• “AI is a fundamental risk to the existence of human civilization.” — Elon Musk, CEO, Tesla and SpaceX
• “The rise of powerful AI will either be the best or the worst thing ever to happen to humanity. We do not yet know which.” — Stephen Hawking, late University of Cambridge professor and physicist
The better question is, what hasn’t been said?
Call it machine learning, deep learning, large language modeling or simply computerized automation. AI is either gushing or seeping into the workplace, depending on the industry. According to a January 2025 report by labor analysis company Lightcast, the number of generative AI-specific job postings has risen 15,625% over three years. There were 156 times as many generative AI postings in 2024 than there were in 2021.
Lightcast’s July 2025 report, entitled “Beyond the Buzz: Developing the AI Skills Employers Actually Need,” shows “explosive” growth in generative AI roles, with an 800% increase in these roles across non-technical industries since 2022. The report ties this increase to companies’ bottom lines, revealing that jobs with AI skills are fetching 28% higher salaries than those without AI skill requirements.
But it’s not just labor analysts who are talking about AI. Those working directly with entrepreneurs and business leaders, like Brand Strategist and Business Growth Advisor Amelia Ellenstein, are hearing the AI “buzz” regularly.
Ellenstein, who has clients across the country, said she’s hearing AI-related questions from all industries. “My clients are asking, ‘how does this challenge or enhance our personal vision for the future?’ she said.
“It’s a real indicator of how important the topic is,” Ellenstein went on to say. “AI represents a huge shift in our culture and the way we work. It’s a culture and strategy conversation.”
Five truths about AI that businesses can’t ignore
There are many things AI isn’t. It isn’t plug-and-play technology. It isn’t magic. It isn’t the solution for every business challenge. It isn’t appropriate for every business task.
But, based on recent research, there are five truths about AI that businesses should understand.
1. AI jobs are not all technical jobs.
Lightcast’s July 2025 AI report examines what it calls “AI in the real world,” drawing from millions of U.S. job postings that reveal the AI skills employers are seeking. The results reveal how “AI has moved from Silicon Valley curiosity to mainstream business necessity,” the report states.
Not long ago, most people would assume a job posting mentioning AI would also include “information technology,” “developer” or “data.” But that’s not the case anymore. Lightcast’s report indicates that most job postings seeking AI skills are non-technical, with AI skills up 800% for jobs outside computer science and information technology.
In 2019, 61% of postings requiring AI skills were in technical fields. That percentage has steadily declined since 2022, just before ChatGPT launched, with 49% of AI postings in technical fields and 51% in non-technical occupations for 2024.
The percentage of job postings mentioning AI has grown across all career fields. In 2016, only five fields mentioned AI skills in at least 1% of postings. In 2024, that number is 13. Top job areas that required AI skills in 2024 included information technology and computer science, with 14% of postings mentioning AI, marketing and public relations postings at just over 8%, science at research at just over 6%, and social analysis and planning at just over 4%. These percentages are low, but year-over-year comparisons show significant growth. For example, marketing and public relations percentages have doubled since 2020. Customer service, finance and education grew from sub-1% in 2020 to just over 1% in 2024.
2. Different AI skills are required in different industries.
Each career area is in a different stage of AI adoption, but the careers also require different AI skills. For example, autonomous driving skills appeared in 40% of AI job postings for transportation positions, and robotics capabilities appeared in 24% of maintenance positions. Science and research roles demanded a wider range of AI skills, with machine learning in 60% of these postings, neural networks in 17% and visual image recognition in 15%. Generative AI were highest for postings in marketing and public relations, and in design, media and writing.
Lightcast’s report also points out that job postings requiring AI skills asked for a set of distinctly human skills alongside AI, like communication, management, leadership, writing and problem solving. “Human skills dominate even in roles explicitly seeking AI expertise, challenging the assumption that AI jobs are purely technical endeavors,” the report states.
3. Low supply and high demand for AI skills equal high labor costs.
Cole Napper, Lightcast vice president of research, pointed out that companies are paying a premium for skilled AI labor. Lightcast’s report shows jobs with one AI skill requirement offering 28% higher salaries than those without AI requirements, and those with two or more offering 43% higher pay. That brings the cost of labor up by $18,000 annually for an employee with just one AI skill, he said.
Ellenstein, who helps company leaders grow their businesses, said she’s seeing this demand play out among her clients.
“There’s a lot of digital information about AI, but not a lot of people you can reach out and engage with,” she said. “Leaders are desperate to find people with expertise to help them.”
4. As companies look for ways to manage the cost of AI expertise, internal AI training and development will become attractive.
Napper said he predicts a “huge pendulum swing toward internal development” of AI skills. As demand for skills continues to increase and companies feel the pain of the price tag for those skills, they’ll move toward developing AI experts within their companies and using them to train other employees.
Ellenstein said companies may also look to develop their own AI systems. Companies are not only seeking to leverage AI as a tool that fits their vision but also trying to figure out how they can keep it special, she said. Most businesses don’t want an “off-the-shelf” tool. They want a proprietary system.
5. Getting a return on investment for AI depends on strategic use of the tool.
A September 2025 Harvard Business Review article references a 2025 MIT Media Lab report, which found that 95% of organizations see no measurable return on their investment in AI. Why? Because of workslop, the article says, defining workslop as “AI generated work content that masquerades as good work, but lacks the substance to meaningfully advance a given task.”
The article also references a study by BetterUp Labs, in collaboration with Stanford Media Lab. The study surveyed 1,150 U.S. workers, and 40% of them reported receiving workslop in the last month. For example, a finance worker creates a presentation detailing quarterly revenues and expenses. When he sends it to his co-worker for review before the presentation, the co-worker notices the verbiage is awkward and key numbers are missing. Does that co-worker ask if the presentation was AI-generated? Notify the boss? Rewrite the presentation?
Beyond reducing productivity by necessitating rework, workslop fosters distrust and hampers teamwork. About half of the people surveyed said they viewed those who sent them workslop as less creative, capable and reliable than they did before, 42% said they saw them as less trustworthy, and 37% saw them as less intelligent. One-third reported being less likely to want to work with the workslop sender again.
Workslop also led to 34% of the people surveyed notifying teammates or managers about the incident, which could erode trust between the sender and receiver.
The lesson here is that indiscriminate use of AI reduces productivity, damages work relationships and inhibits collaboration. “Gen AI is not appropriate for all tasks, nor can it read minds,” the article states. “AI can positively transform some aspects of work, but it still requires thoughtful guidance and feedback from workers in order to produce useful outputs on complex or ambiguous work.”
Productive use of AI depends on setting guardrails for how and when employees should use it, the article states. Elizabeth Williams, principal of HR solutions for Rehmann, seconds that notion. “Make sure your employees are trained on what things AI can do and what it can’t do,” she said.
It’s important to acknowledge and address AI’s shortfalls, Williams noted. One is that it can “hallucinate” by inventing facts when it doesn’t have answers, she said. “AI is only as good as what’s out there. It’s not as it thinks it is. It pulls from the best answer it knows.”
The antidote to those hallucinations is human discernment. People can use AI to generate ideas and provide guidance, but they must always check its accuracy and tweak its output as necessary.
The Harvard Review article describes two types of AI users — pilots and passengers. Pilots have high agency and high optimism, and they seek to use AI to enhance their creativity. Passengers, on the other hand, have low agency and optimism, and they use it to avoid doing work.
Leaders who ask employees to operate under a “pilot” mindset for AI business use and model that mindset themselves are likely to get a better return on their AI investment, the article states.
Williams noted other AI cautions, including sharing company data on an open system and discrimination. There have been lawsuits from AI systems that have discriminated against certain groups, she said. For example, a beta tested system screened out females and gave them a lower rating because the engineering field is male dominated. If you don’t understand the criteria your AI system is using to choose candidates, you can not only miss certain sources of talent but also expose yourself to legal recourse.
Business advice for AI use
Given these facts about AI, how should companies move forward? There’s not a one-size-fits-all answer, because it depends on the industry and the individual company.
Napper advised companies investigate how AI can advance their business strategies and then make decisions about it accordingly. Even within the same industry, those decisions will vary, he noted. For example, some companies might go all in on AI use for customer service while some might decide to only use humans as a market differentiator. It all depends on the business’s strategy and customer base.
Ellenstein’s advice echoed Napper’s sentiment. “Lean heavily into the vision and values of the company,” she said. “Then take action in ways that align with the vision and values.
“Companies that rely on their unique vision and values to guide decision-making can clearly articulate those decisions and policies in a sincere and natural way.”
Practical advice from Williams included developing an internal AI committee and working with the information technology department or external IT consultants to understand how it works. Ask IT to “give us a way to play in the sandbox,” she said. Finally, develop an AI policy and roll it out to employees.
The future of AI
Going back to our original quotations, what’s the value of AI for business use? Is it a blessing or a curse? Of course, it depends on who you ask, but many see it as a useful business tool for those who apply it strategically.
“I’m excited about the opportunity to innovate and the ability to be creative,” said Ellenstein. She said she hopes people will use it as a tool to stimulate thinking and brainstorm. “It’s a tool, and we get to choose how we use it.”
One final thought from Napper puts AI for businesses into perspective. Reading many of the AI quotes at the beginning of this article likely causes one’s heart to beat a little faster in fear. However, while Lightcast statistics do show exponential growth in demand for AI skills, there’s still a long way to go before AI is employed in every industry. “Maybe some of the fear is overblown,” he said. “AI skills have not pervaded us as much as people expect. It’s moving slower than the headlines indicate, and that’s a good thing. It gives people space to breathe.”
