Spellman High Voltage Electronics Corporation

Stefan Dumlao brings a data-driven perspective to the changing relationship between people, technology, and work. His focus on people data, analytics, and insights places HR at the center of questions around how jobs are changing, how employees use AI, and how organizations should rethink talent management as technology reshapes work.

AI Is Changing Jobs from Within

Predictions that AI would rapidly eliminate large numbers of white-collar jobs have been widespread. In January 2026, Anthropic chief executive Dario Amodei restated his forecast that AI could displace half of all entry-level white-collar jobs within one to five years. In 2025, OpenAI chief executive Sam Altman predicted that entire categories of jobs would disappear, while Ford chief executive Jim Farley forecast that AI would eliminate half of U.S. white-collar jobs. By May 2026, Altman said he was delighted to be wrong. In August, Meta chief executive Mark Zuckerberg argued that people will need to adapt to changes in how they work without that necessarily resulting in fewer jobs overall.

Current labor-market data points in the same direction. AI is reshaping white-collar work, but much of that change is happening within jobs rather than appearing as broad unemployment. The question for HR is not only which jobs might disappear, but how the work inside those jobs is changing.

Employees are already reshaping their roles, a practice I-O psychologists call job crafting. Analyzing more than 800,000 messages from U.S. ChatGPT users, OpenAI found that 43.5 percent of occupation-specific messages involved tasks associated with another occupation. Among human resources workers, that figure reached 69 percent, with requests frequently involving marketing, engineering and finance. Jobs as performed may be drifting away from jobs as documented, while most HR systems still begin with documented roles, responsibilities and skills.

Defining AI Skills for the New Workplace

Not all AI use creates value in the same way. Efficiency AI uses AI to perform existing work faster, creating value in routine, well-defined tasks where results are easy to evaluate. Opportunity AI uses AI to take on work that would otherwise remain out of reach because of limited time, budget or people. Much of the crossover identified in OpenAI’s data appears to reflect this second category, with workers using AI to supplement their expertise with knowledge outside their primary domain.

“The more consequential the work, the more an AI system needs a human in the loop—and the less HR can afford to be left out.”

Both forms of use place greater demands on employees. Individual judgment and accountability remain essential. Researchers at Glean distinguish botsitting, the untracked labor involved in making AI usable, from shipping AI-generated work that employees have not reviewed, do not fully understand or could not defend if challenged.

Good AI practice is visible in behaviors that organizations can observe and develop. Employees need to explain and defend AI-assisted work, scope requests clearly, recognize when a session has gone off course and start again when necessary. They also need to understand which model and level of effort fit the task.

These behaviors are increasingly relevant to hiring. A spring 2026 National Association of Colleges and Employers survey found that more than one-third of entry-level jobs require AI skills, nearly triple the share reported in fall 2025. IBM has also rewritten roles around AI fluency while increasing entry-level hiring for software development and other positions being transformed by AI.

For HR, the practical question is what AI skills should mean on a job description. Familiarity with a particular tool is unlikely to be enough; the behaviors described above provide a stronger foundation.

Keeping Job Architectures Current

Understanding these skills begins with knowing what a job involves. Traditional job analysis relies on SME interviews, observations and task ratings, processes designed for roles that remain stable long enough to be analyzed.

Rapid AI adoption challenges that assumption. HR needs new ways to capture, analyze and maintain job and skills architectures as work changes. Systems could allow employees and managers to report task changes as they occur. Hiring interview notes and application metadata could also identify signals that a role is evolving and prompt updates to task statements.

Performance management needs similar attention. The volume of an employee’s AI use reveals little about the quality of that use. Effective experimentation also includes attempts that do not succeed. Yet Microsoft’s 2026 Work Trend Index found that only 13 percent of AI users said they are rewarded for reinventing their work with AI, even when results are not met.

Organizations that encourage AI experimentation but recognize none of that effort risk relying only on intrinsically motivated individuals. Documented trials, shared workflows and after-action reviews should become visible parts of role expectations, performance reviews and advancement.

Putting HR at the Center of AI Strategy

HR also needs a stronger role in organizational AI strategy. SHRM’s State of AI in HR 2026 found that 52 percent of organizations do not involve HR in AI strategy or vision.

Many questions surrounding AI at work are fundamentally questions about people: how employees interact with AI tools, how teams of people and agents should be designed, what distinguishes appropriate use from misuse, how adoption should be encouraged and how performance should be evaluated.

The more consequential the work, the more an AI system needs a human in the loop—and the less HR can afford to be left out.