Two-Thirds of Employers Added AI Skills to Non-Technical Jobs. Most Changed Neither Title Nor Pay

Most job redesigns used to come with paperwork: a new title, a revised description, a pay review, and a conversation with a manager. The AI version is skipping all of it. According to Oyster's 2026 Global AI Talent Report, released October 14, AI skill requirements have tripled since 2022, and the new expectations are increasingly landing on people whose jobs have never been described as technical. The work has changed. The job, as the organization records it, often has not.

The research combines an analysis of roughly 15,000 jobs filled through Oyster's global employment platform between 2022 and the first half of 2026 with a survey of more than 1,000 HR and people leaders in the United States, United Kingdom, Canada, and Europe. Because the job data covers only roles filled through Oyster, whose customers already hire across borders, the platform figures are best read as a leading indicator rather than a census. The survey findings, though, describe a pattern most HR teams will recognize.

The Job Description Is Now a Historical Document

In the full survey results published by TecHR Series, 95% of HR professionals said employees are using AI tools every day in roles whose official job descriptions never mention AI. Two-thirds said their companies have added AI skill requirements to non-technical roles in functions like marketing, HR, and operations. Oyster's platform data shows how far that has spread: 17% of marketing coordinator roles, 15% of recruiter roles, and 10% of operations coordinator roles now require AI skills, and nine in ten AI-requiring roles do not mention AI in the job title at all.

The administrative follow-through is where the pattern breaks. Of the companies that added AI requirements, 46% have not formally communicated the change to the people affected. Among companies that added AI requirements to existing roles in the past 12 months, 62% changed neither the title nor the pay. And when an employee automates part of their own job, only 38% of HR professionals said that person would be rewarded, while 24% said they would simply be given more work and 26% said nothing would change.

"The people making workforce decisions are often the furthest from the tools doing the work." – Hadi Moussa, CEO, Oyster

That gap between expectation and recognition has a trust cost. A Boston Consulting Group survey of 645 CEOs and 500 nonexecutive employees, also released October 14, found that 40% of employees say their CEO has not been transparent enough about how AI will affect their roles and careers, while only 15% of CEOs worry that their AI approach could erode trust. Changing what a job requires without saying so is exactly the kind of action employees read as a lack of candor.

Training Is the Preferred Fix, but Not the Delivered One

The encouraging finding in Oyster's data is that most employers want to build rather than buy. When AI skills are missing, 53% of companies said they prefer to train existing employees, 31% combine training with hiring, and only 14% mainly hire externally. Yet employees are not experiencing much of that training. In a Visier survey of 1,000 US full-time employees, 54% said their roles or careers had changed significantly because of AI in the past two years, but only 21% strongly agreed that their employer provides the training, resources, and support they need to adapt. More AI training and upskilling was the top thing employees said would build their confidence.

Hiring signals are muddled too. Oyster found that 68% of companies ask candidates what they have built with AI, but 53% of those companies say the answer does not affect the hiring decision or the offer. Organizations are asking about AI capability, requiring it in the role, and declining to price it, which leaves both candidates and incumbents guessing what the skill is actually worth.

When the Talent Search Crosses a Border

Companies that cannot find AI-ready people at home are increasingly looking abroad. Oyster reports that companies prioritizing external AI hiring are nearly twice as likely to have recruited internationally to fill the gap: 34% have done so, against 18% of companies that prioritize training. Nearly eight in ten companies have never hired internationally for a domestic AI shortage, but 26% of that group are now considering it. Demand is not evenly spread, either. On Oyster's platform, 21% of roles in Switzerland, 20% in Poland, and 19% in Portugal require AI skills, compared with 9% in the United Kingdom.

That is where a skills story becomes a compliance story. Some 29% of companies using cross-border AI contractors said they are not confident those engagements are correctly classified under local labor law, or have never considered the question. The doubt rises to 31% among companies that staff AI roles mainly with contractors, compared with 20% among those that mainly use full-time employees. Experience helps: 87% of companies that have hired internationally and use international contractors are confident in their classification, against 65% of companies that have never hired abroad. The companies newest to global hiring are the least certain they are doing it legally.

Oyster also found that 21% of HR professionals have seen an AI tool or self-service HR system miss something important in a high-stakes situation, such as a termination or a leave request. Automation is reshaping the people function itself, and the same documentation discipline applies inside HR as everywhere else.

The Playbook for HR Leaders

• Audit job descriptions against actual work. Start with the functions Oyster flags, marketing, recruiting, and operations, and document where AI use is now an expectation rather than an option.
• Tell people when their job changes. With 46% of companies skipping formal communication, a written note explaining new expectations, the reason for them, and the support available closes a trust gap BCG puts at 40% of employees.
• Price the requirement or remove it. If a role now requires AI skills, review its title, level, and pay band. Leaving 62% of redesigned roles untouched invites both attrition and pay equity questions.
• Match the training to the preference. Most employers say they would rather train than hire, so fund the training before adding the requirement, not after, and measure whether employees feel supported.
• Reward automation instead of punishing it. If employees who automate part of their job only get more work, they will stop telling HR what they have built. Define how efficiency gains are recognized.
• Classify cross-border AI talent before engaging it. Review every international contractor arrangement against local employment tests, and consider employer-of-record or full-time models where control and integration look like employment.

AI is redrawing jobs one task at a time, and most organizations are letting the redrawing happen without updating the map. The result is employees doing work their titles do not describe, for pay that does not reflect it, under arrangements that may not survive a labor audit abroad. HR does not need to slow the change down. It needs to write it down, pay for it, and tell people it is happening.