If 2023 was the year of shock and 2024 was the year of experimentation, 2025 marks the moment the corporate world finally accepted that artificial intelligence is not just a novelty—it is the new infrastructure of work. According to Justina Nixon-Saintil, IBM’s Vice President of Corporate Social Responsibility and Chief Impact Officer, the conversation has fundamentally shifted from fascination to urgent integration.

In a recent interview with Fortune at IBM’s gleaming new flagship office at the southeastern corner of Madison Square in New York City, Nixon-Saintil pinpointed 2025 as the specific timeframe when “the penny dropped” for global industries.

“It was sometime this year where I feel like the conversation changed from, ‘Oh, there are these really cool virtual assistants’ … to, ‘Oh, wow, companies are investing in this in a big way, and this is actually happening, and it’s transforming work,'” Nixon-Saintil said. She noted that as recently as July, industry panels were still debating whether to halt development. Her response to the skeptics was blunt: “No, AI disruption is not stopping … there’s no stopping what’s happening. We have to quickly move to action.”

The rise of ‘power skills’

As the inevitability of AI has set in, Nixon-Saintil described seeing how the anxiety regarding job displacement has evolved. The VP, who works extensively with colleges as a top-level liaison from the corporate sector, said the definition of “knowing how to use AI” is changing. It is no longer enough to possess technical proficiency; the workforce now demands what Nixon-Saintil calls “power skills” or “soft skills”—specifically, the ability to apply human oversight to algorithmic output.

“You can’t hire a college student now to just come in and create a spreadsheet,” Nixon-Saintil explained, emphasizing that rote tasks are rapidly being automated. Citing IBM’s internal research, she said the premium is now on “understanding the work enough to make the right judgment calls, critical thinking, all of those other skills … not the tech skills.” While technical skills are a baseline requirement, the ability to exercise judgment over an AI agent’s work is becoming the true differentiator in employability.

The IBM executive’s remarks echoed those from Kelly Monahan, managing director of the Upwork Research Institute, who told Fortune in September that “humans are coming back into the loop.” Upwork, a global work marketplace for freelancers, was seeing “human skills coming into premium,” she added. This was in part a response to what Deutsche Bank termed “the summer AI turned ugly” as hallucinations proved too pervasive for companies to simply plug AI into their processes. Hence the importance of critical thinking identified by Nixon-Saintil: AI isn’t going away, but it isn’t a cure-all, either, so finding the top talent to use it is more important than ever.

Global consulting firm Protiviti highlighted talent as one of the major risks for executives heading into 2026, with a panel of experts voicing concern about a skills gap in this regard bigger than any they can remember. As recently reported by Fortune, North Carolina State University’s Dr. Mark Beasley, who has been affiliated with the annual Protiviti risk survey for 14 years, told reporters that “We need to start thinking strategically. How can we create strategic thinkers, critical thinkers?” When asked if he was worried about a “thinking gap,” he said: “Yes. As a university professor, yes, I am.”

It’s an urgent question with the U.S. still mired in a “low-hire, low-fire” economy, with an elevated unemployment rate for recent college graduates and a plunge in entry-level jobs as executives are mystified about the talent question. The Business Roundtable CEO survey for the fourth quarter of 2025 marked the third consecutive period in which more CEOs anticipated their company’s employment would decrease rather than increase. Julia Coronado of MacroPolicy Perspectives told reporters at the Protiviti luncheon that the missing entry-level jobs will soon lead to a crisis in middle management if this goes unaddressed: “If AI is sort of replacing the entry level typical positions, and I need people sort of in the middle, how do I prepare the future middle if I don’t give them that ability at the base?”

Exploding demand: The Saudi case study

From her standpoint, Nixon-Saintil is optimistic, highlighting the work IBM has done to upskill workers at scale already. The urgency to acquire skills has led to an explosion in demand for training that has outpaced even aggressive corporate targets, she told Fortune. She highlighted IBM’s announcement in December 2025 that it had vastly exceeded its upskilling goals in Saudi Arabia.

“We had committed a couple of years ago to upscale 100,000 people in Saudi Arabia” by 2027 as part of IBM’s work for Vision 2030, she said, revealing that IBM just passed the 500,000 mark, a year ahead of schedule. This milestone, achieved in collaboration with the Saudi Ministry of Communication and Information Technology, underscores how nations are racing to align their workforces with a “knowledge-based, innovation-driven economy.”

Regarding the universities she works with as part of IBM SkillsBuild, Nixon-Saintil said they had largely abandoned the futile effort to ban generative AI, realizing that “students are going to use it no matter what.” She said she’s seeing a shift toward teaching responsible-use guidance and integrating AI fundamental courses, recognizing that students are already arriving with baseline familiarity of AI tools. The focus has shifted to preparing a generation of students who are entering higher education as “AI natives.”

The VP revealed that she has two college-age kids, and she told Fortune that she gives them the same advice she gives all the students she meets: make sure that you’re not using AI instead of learning a topic; use it to tutor you or go deeper on a topic, but make sure you’re not using it to create all of the output that you need. She said she sees college kids using AI similar to how workers are: “It helps you by giving you a jumpstart or it helps you by analyzing a lot of data and giving you some ideas, right? And in some cases it will create something for you, but you also have to have that judgment call. You have to review it. You have to make sure your voice is in it, and I think that’s important for people to understand.”