Laid Off by AI or Paid More? It Depends on This
Introduction
Something strange is happening in the job market right now. Workers who regularly use AI tools are pulling in salaries up to 28% higher than peers doing similar work without them — while Goldman Sachs analysts simultaneously warn that workers displaced by automation may face nearly a decade of lower wages before recovering.
Both data points are accurate. The AI impact on salaries is not a single, clean story. It is two stories unfolding at the same time, and which one you are living depends almost entirely on whether AI sees you as competition or as a co-pilot.
The divide is sharpening. And the line separating the two groups is more learnable than the doom-and-gloom coverage suggests.
The Numbers Tell a Story — Just Not the One You Expect
Lightcast's 2025 research landed with a striking finding: job postings that require AI skills offer salaries roughly $18,000 higher on average than comparable roles without that requirement. That translates to a 28% premium in some sectors. Finance analysts working with AI forecasting tools, marketing managers who can run AI-driven campaign optimization, healthcare administrators managing AI diagnostic workflows — all of them are being paid more.
Then there is the Goldman Sachs analysis pointing the other direction. Workers displaced by AI-driven automation historically take years, sometimes close to a decade, to recover their previous earnings. The path back is rarely clean or linear.
These findings are not contradictory. They are describing two completely different populations. The question worth asking is which population you belong to — and whether you can choose.
Why Tech Sector Layoffs Miss the Bigger Picture
AI layoffs in the tech sector have been real and significant. Major companies cut thousands of roles across customer support, content moderation, junior software development, and data annotation. The press treated it as a warning shot aimed at every knowledge worker. That reaction was understandable.
But here is what that narrative consistently misses: AI skills are not staying inside tech. They are dispersing rapidly into every sector.
The Lightcast research makes this explicit. The fastest AI-driven salary growth is showing up in manufacturing, logistics, finance, and healthcare — not at big tech companies. Mid-sized firms that have never hired a machine learning engineer are hiring operations managers who can work with AI scheduling tools. Regional hospital systems are paying more to administrators who understand AI-assisted billing workflows. Construction companies are offering retention bonuses to project managers who use AI cost forecasting software.
Equating AI layoffs in the tech sector with AI being broadly bad for salaries is a category error. The disruption is real. But automation wage growth is also real — it just flows to the people who adapt rather than the ones who wait.
A concrete example worth sitting with: a supply chain analyst at a consumer goods company who began using AI-driven demand forecasting tools in 2024 reported a 22% salary increase when changing employers the following year. Her title did not change. Her core domain expertise did not change. What changed was the multiplier she brought to that expertise.
The Skill Gap Is More Specific Than Just Learning to Code
When people talk about acquiring AI skills, the conversation usually collapses into two buckets: build AI systems (machine learning, model training, fine-tuning), or use AI tools. The second category is orders of magnitude larger and far more accessible.
In practice, what actually happens is that most AI-linked salary gains do not require any coding knowledge. They require fluency. A financial analyst who knows how to use AI tools for research synthesis and report drafting does not need to understand attention mechanisms or transformer architecture. She needs to know how to construct clear prompts, how to verify outputs against ground truth, and how to integrate AI-generated work into a process her team and her clients can trust.
Many practitioners find that the steepest part of the learning curve is not technical — it is behavioral. The instinct to just handle it yourself, the quiet suspicion that AI output is probably wrong, the friction of changing a workflow that already works adequately. These are not irrational responses. They are just expensive ones, compounded over time.
Workers who push through that friction and genuinely build AI into their daily process often report an unexpected side effect: not just speed, but quality. Less time on assembly work means more time on judgment calls, synthesis, and communication. Those happen to be exactly the skills that justify higher compensation.
That is the AI productivity gains mechanism in practice. Not replacement. Augmentation of the parts of work that are already most human.
The Counterargument: Is the Salary Premium Just Temporary?
Some argue that the current wage premium for AI-fluent workers is a transitional artifact. Give it three years, the thinking goes, and AI systems will be capable enough to handle the prompting, the output verification, the workflow integration — all of it. The loop closes. Human salaries fall everywhere.
It is a coherent argument. But it misses something structurally important.
AI systems require organizational trust, not just technical capability. In regulated industries — healthcare, law, financial services, government contracting — there are humans in the loop by design and by regulation. Someone needs to audit AI decisions, explain them to stakeholders, catch errors before they become liability, and accept accountability when things go wrong. That role does not shrink as AI becomes more capable. It frequently grows more critical. A highly confident AI system that produces fluent-sounding errors is considerably more dangerous than a slow human who catches them.
The jobs most at risk are not the ones requiring judgment, complex communication, and accountability. Those have historically been the hardest jobs to automate, and they remain so. The roles under real pressure are execution-heavy, pattern-repetitive positions that were already fairly compensated rather than highly compensated.
The future of work with AI is a bifurcation, not a uniform slide. Workers who develop the capacity to supervise, interpret, and deploy AI systems do better. Workers whose primary value was execution speed and volume do worse. That is uncomfortable to say plainly. But it is more useful than false reassurance.
Where Non-Tech Salary Increases Are Actually Showing Up
A practical map of where AI-linked salary gains are occurring outside tech tells you something useful about where to direct your attention.
Healthcare administrators at large hospital systems using AI for scheduling, billing optimization, and compliance reporting are earning more — partly because the efficiency gains are measurable and institutions want to retain the people who produced them.
Legal professionals who use AI-assisted contract review are handling higher caseloads without proportional staff increases. Billing rates are holding or rising because the bottleneck shifted from document review to strategic analysis — and clients are still paying for the latter.
Supply chain managers working with AI demand forecasting are being actively recruited with retention packages. Multiple major retailers in 2024 lost supply chain managers to competitors specifically because of AI tool proficiency gaps.
The consistent pattern: AI fluency combined with deep domain expertise creates a multiplier that neither element produces alone. Plenty of people with AI tool knowledge but no domain depth are struggling to translate that knowledge into actual earnings. The salary premium lives in the combination, not in either side of it.
Honestly, that is a more encouraging finding than the raw headline numbers suggest. You do not need to become a technologist. You need to become the most AI-capable person in your specific field. That is a narrower and more achievable target for most working professionals.
A Practical Framework for Figuring Out Where You Stand
Rather than asking whether AI will affect your salary in the abstract, three sharper questions cut closer to the answer.
First, is your role primarily about execution or judgment? Roles built around executing defined processes — data entry, routine reporting, standard customer service scripts — face more pressure than roles centered on interpretation, decision-making, and complex communication. This is not a comfortable question, but it is an honest one.
Second, does your industry carry regulatory or accountability requirements? Highly regulated sectors maintain human oversight requirements that are not going away quickly. If your field requires professional licensing, fiduciary duty, or legal accountability for decisions, you have structural protection that many pure-tech roles lack.
Third, are you visible when AI projects come up at your organization? This last one is underrated. Workers who ended up on the wage-growth side of AI's impact on salaries often share a common trait: they were early testers, they gave feedback on new tools, they became the person colleagues asked for help. That informal expertise accumulates into formal recognition faster than most people expect. It also makes you harder to replace, which is precisely the point.
None of these are guarantees. The macro pressures — AI layoffs in specific sectors, extended earnings disruption for displaced workers — are real at the population level. But at the individual level, these three questions frequently explain more of the outcome than the industry-wide trend does.
Your Next Move Matters More Than the Headlines
The AI impact on salaries is not a single verdict delivered from on high. It is a split outcome playing out right now, with the dividing line running through skill adoption rather than job categories.
Workers who build AI fluency early, inside fields they already know well, tend to find themselves on the wage-growth side of this divide. Workers who treat AI as someone else's problem tend to find themselves on the displacement side.
That is the actual tension. Not man versus machine. Not tech versus everyone else. It is early adopters versus late adopters, in every field simultaneously.
The cost of being late is rising. The cost of starting now is a few hours of discomfort learning tools that will be standard in your industry within two years anyway. That trade-off is not even close.
Pick one AI tool relevant to your actual work. Get genuinely proficient at it, not just passably familiar. Then watch what it does to the quality — and eventually the price — of what you produce.