AI is creating jobs faster than it eliminates them, but the roles are different from the ones disappearing

The jobs AI creates are not the same as the jobs it replaces. A factory that automates assembly does not rehire those workers as "automation technicians" — but the company that builds the automation software hires dozens of engineers, trainers, and support staff. The net effect across the economy is job growth, but it happens in different places, requires different skills, and pays differently than what came before.

Right now, AI is creating roles that did not exist five years ago: prompt engineers who write instructions for AI systems, AI trainers who teach models to behave correctly, and data annotators who label images and text so AI can learn from them. It is also expanding older roles — data scientists, software engineers, and business analysts all have more demand and higher pay because companies need people who understand how to use AI tools. At the same time, it is shrinking demand for data entry clerks, some customer service representatives, and junior-level coding jobs that AI can now handle.

Key Takeaways

  • AI creates jobs in software development, data work, and training roles, but eliminates routine jobs like data entry and some customer service positions.
  • The jobs that remain often require people to work alongside AI tools rather than replace the tools themselves — knowing how to use AI is becoming a baseline skill.
  • Wages are rising for technical roles that involve AI, but falling or stagnating for routine work that AI can automate.
  • The transition is uneven: some industries and regions are adding jobs while others are losing them, and workers in declining roles often cannot straightforward move to new ones without retraining.

Jobs AI is creating right now

Prompt engineering is the most visible new role. A prompt engineer writes detailed instructions for AI systems to produce specific outputs — for example, asking an AI image generator to create a product photo with exact lighting and background, or asking a language model to summarize a legal document in a particular style. Companies hire prompt engineers to handle tasks that require consistency and quality control. The role typically requires a college degree and pays between $80,000 and $150,000 depending on experience and location.

AI trainers work with machine learning models to improve their accuracy and behavior. They review outputs, flag errors, and provide feedback that the model learns from. This role is more common than most people realize — companies like OpenAI, Google, and Meta employ thousands of trainers. It usually requires a high school diploma and attention to detail, though some positions prefer college education. Pay ranges from $35,000 to $70,000 annually.

Data annotators label images, text, audio, and video so that AI models can learn to recognize patterns. An annotator might draw boxes around cars in thousands of photos, or mark which parts of a customer service call were handled well. This work is often contract-based and pays $20 to $40 per hour, though some positions are remote and flexible.

AI safety and ethics roles are growing as companies face pressure to build responsible systems. These positions involve testing AI for bias, ensuring it does not produce harmful outputs, and documenting how it makes decisions. They typically require a college degree in computer science, philosophy, policy, or a related field, and pay $90,000 to $180,000.

Jobs that are expanding because of AI

Existing technical roles are growing and paying more because AI tools have made them more valuable. A data scientist who used to spend weeks cleaning data can now use AI to do that in days, freeing them to focus on strategy and interpretation. Companies are hiring more data scientists and paying them more — median salary is now around $120,000, up from $100,000 five years ago.

Software engineers are in higher demand because building AI systems requires more code, not less. AI tools like GitHub Copilot help engineers write code faster, but companies are hiring more engineers to build the systems that use those tools. Salaries for software engineers have remained stable or risen, typically $110,000 to $200,000 depending on experience and location.

Business analysts and product managers are increasingly needed to figure out where AI can actually help a company and how to measure whether it is working. These roles sit between the technical team and the business side, and they require understanding both how AI works and what the company is trying to achieve. Salaries range from $80,000 to $160,000.

Trainers and change management specialists help employees learn to use new AI tools. When a company adopts an AI system, someone has to teach people how to use it, what it can and cannot do, and how it changes their workflow. These roles are often filled by people who already work in the company but are given new titles and responsibilities.

Jobs that are shrinking or disappearing

Data entry is the clearest example. AI and automation can now read documents, extract information, and enter it into databases faster and cheaper than humans. Companies still hire data entry workers, but they hire far fewer than they did ten years ago. Wages for this work have stagnated or fallen — positions that paid $30,000 in 2015 often pay $28,000 today.

Customer service representatives handling routine questions are declining in number as chatbots and AI systems handle first-contact support. A company might have 100 customer service agents today and 60 in five years, with the remaining 40 handling complex issues that require human judgment. Wages in customer service have not kept pace with inflation.

Junior software developers and junior writers are facing more competition from AI tools that can generate basic code and content. A company that would have hired three junior developers to build a straightforward website might now hire one senior developer and use AI tools for the routine parts. This does not mean junior roles are disappearing entirely, but the number of entry-level positions is shrinking.

Transcriptionists, proofreaders, and some graphic designers are seeing demand decline as AI tools improve at these tasks. A transcriptionist might now spend their time editing AI-generated transcripts rather than creating them from scratch, which is faster but pays less.

How AI is changing the jobs that remain

Even jobs that are not disappearing are changing. Most roles now involve using AI tools as part of the work. A marketer uses AI to analyze customer data and suggest campaign ideas. A lawyer uses AI to search case law and draft contracts. A doctor uses AI to interpret imaging scans. These are not new jobs — they are the same jobs with different tools.

This shift creates a skills gap. Workers who learned their trade before AI tools existed often struggle to adopt them, while younger workers who grew up with AI expect to use it. Companies are investing in retraining programs, but they are not keeping pace with the speed of change. A 50-year-old accountant who learned to do their job by hand and then adapted to spreadsheets now has to learn AI tools — and if they do not, their job becomes less find.

The jobs that are safest are ones that require judgment, creativity, or human connection — things AI still cannot do well. A therapist, a manager, a teacher, a nurse, and a skilled tradesperson are less vulnerable to automation than a data entry clerk or a junior analyst. But even these roles are changing as AI tools become part of the work.

The uneven impact across industries and regions

AI is not creating jobs evenly. Tech hubs like San Francisco, Seattle, and Austin are adding AI-related jobs quickly. Manufacturing regions that relied on routine factory work are losing jobs faster than new ones appear. A town where the main employer is a call center faces a real threat if that company automates customer service.

Some industries are adding jobs while others shrink. Software, finance, healthcare, and education are all hiring for AI-related roles. Retail, manufacturing, and transportation are losing jobs to automation. A person whose job is disappearing in one industry cannot straightforward move to a growing industry without retraining, and retraining takes time and money that not everyone has.

This is why the overall statistic — "AI creates more jobs than it destroys" — can be true while also being unhelpful for a specific person in a specific place. If you are a customer service representative in a town with one major employer, the fact that AI is creating jobs in San Francisco does not help you.

What skills matter for AI-era jobs

The jobs that are most find involve skills that AI cannot easily replicate. These include critical thinking (deciding whether an AI output is correct), communication (explaining technical concepts to non-technical people), and domain informed (knowing enough about your field to know what questions to ask).

Technical skills are valuable but not always necessary. You do not need to be a programmer to work with AI. Many of the fastest-growing roles — prompt engineering, AI training, data annotation — require attention to detail and the ability to follow instructions, not advanced degrees. However, learning to code, understanding statistics, or studying data science opens more doors and typically leads to higher pay.

Soft skills are becoming more important, not less. As AI handles routine work, the jobs that remain involve collaboration, problem-solving, and explaining decisions to others. A person who can work well with a team, adapt to new tools, and communicate clearly is more valuable than someone with narrow technical informed.

Frequently Asked Questions

Will AI take all the jobs?

No. AI will eliminate some jobs and create others, but the total number of jobs in the economy is likely to grow. However, the new jobs may not be in the same places or pay the same as the jobs that disappear. A person whose job is eliminated may not be able to move into a new one without retraining.

What is the best job to learn if I want to work with AI?

It depends on your background. If you like coding, software engineering and machine learning engineering are in high demand and pay well. If you prefer non-technical work, prompt engineering, AI training, and product management are growing. If you are just starting out, data annotation or AI training roles require less experience and can lead to better positions later.

Do I need a college degree to get an AI job?

Not always. Data annotation and AI training roles often require only a high school diploma. Prompt engineering is newer and has no standard requirements yet. However, technical roles like software engineering and data science typically require a degree or equivalent experience. Many people learn through online courses, bootcamps, or self-study, though a degree still opens more doors.

Are AI jobs only in tech companies?

No. Banks, hospitals, retailers, manufacturers, and government agencies all hire for AI roles. Any large organization that uses data is likely to need people who understand AI. However, the highest concentration of AI jobs is still in tech companies and tech-heavy regions.

What happens to workers whose jobs are being automated?

Some find new jobs in growing fields, often with retraining. Others move to different companies or industries. Some stay in their field but do different work — for example, a transcriptionist might edit AI-generated transcripts instead of creating them from scratch. Government programs and some companies offer retraining, but coverage is uneven and often insufficient.