What's actually happening to jobs right now

AI is not eliminating jobs wholesale — it is changing what jobs exist and who does them. Some roles are shrinking because software now handles the work. Other roles are growing because companies need people to build, maintain, and work alongside AI systems. The net effect depends on your industry and skill level, and it is uneven across different fields.

The jobs that are growing fastest are not the ones you might expect. It is not just software engineers. Companies are hiring data annotators to label images for training, prompt engineers to write instructions for AI systems, AI trainers to test whether outputs are accurate, and compliance officers to make sure AI systems follow regulations. These are new categories that barely existed five years ago.

At the same time, some jobs are shrinking. Customer service roles that involved answering routine questions are declining in number. Data entry positions are disappearing. Certain kinds of writing work — product descriptions, basic content for websites — is being done by AI instead. But the people who did those jobs do not automatically move into the new ones. That is the real problem.

Key Takeaways

  • Jobs that are growing include AI trainers, data annotators, prompt engineers, and people who manage AI systems in specific industries like healthcare and finance.
  • Jobs that are shrinking include routine customer service, basic data entry, and some writing and design work that AI can now do without human input.
  • The new jobs usually require different skills than the old ones, so someone who did customer service for ten years cannot straightforward move into AI training without learning new things.
  • Growth is fastest in fields where AI is newest — healthcare, law, finance — because those industries are still figuring out how to use it and need people who understand both the technology and the field.

Jobs that are growing because of AI

Data annotators label images, text, audio, or video so that AI systems can learn from examples. A company building a medical AI system might hire someone to mark where tumors appear in thousands of X-rays. A self-driving car company hires people to draw boxes around pedestrians in video footage. The work is repetitive but does not require a degree — many companies hire remotely and train on the job. Pay ranges from $15 to $25 per hour depending on the complexity and the company.

Prompt engineers write instructions for AI systems to get better results. Instead of coding, they experiment with how to phrase requests to ChatGPT, Claude, or similar tools so that the output is more useful. A marketing team might hire someone to write prompts that generate product descriptions faster. A law firm might hire someone to write prompts that help AI summarize contracts. This role is still new and pay varies widely — some companies pay $80,000 to $150,000 per year, while others treat it as a junior role at $40,000 to $60,000.

AI trainers and evaluators test whether AI systems produce correct answers. They run the system through scenarios, check whether the output is accurate, and flag problems. A healthcare company might hire someone to verify that an AI system correctly identifies diseases. A financial services company might hire someone to check whether an AI system gives sound investment information. These roles usually require domain knowledge — you need to understand healthcare or finance, not just AI. Pay is typically $50,000 to $90,000 per year.

AI compliance and ethics roles are growing in regulated industries. Banks, healthcare providers, and insurance companies need people who understand both AI and regulatory requirements. These roles involve auditing AI systems to make sure they do not discriminate, documenting how decisions are made, and ensuring the company follows laws. They usually require a background in the industry plus some AI knowledge. Pay ranges from $70,000 to $130,000 per year.

Jobs that are shrinking because of AI

Routine customer service is declining because chatbots now handle first-contact questions. If you call a bank and ask about your balance or a credit card limit, an AI system answers. If you email a company asking when your order ships, an AI system responds. Companies still need customer service workers, but fewer of them, and the remaining roles focus on complex problems that AI cannot solve. This shift has been happening for years but is accelerating.

Data entry and basic administrative work is disappearing because AI can read documents, extract information, and populate databases. A company that once hired five people to enter invoice data into a system now uses AI to do it and hires one person to check the results. The Bureau of Labor Statistics projects that data entry roles will decline by about 5 percent over the next decade, which is faster than average job growth.

Some writing and design work is shifting. Companies that once hired writers to produce product descriptions, social media posts, or basic web copy now use AI to generate drafts and hire fewer writers to edit them. Graphic designers report that AI tools are handling more routine design work. This does not mean writing and design jobs are disappearing entirely — they are changing. The writers who remain tend to do more strategic work, and the designers who remain tend to do more complex or custom work.

Transcription and translation roles are declining as AI improves. A decade ago, companies hired people to transcribe audio recordings or translate documents. Now AI does much of that work, and humans check the results. The number of transcriptionists in the United States has fallen by roughly 20 percent since 2015, according to Bureau of Labor Statistics data.

Where the growth is fastest

The industries hiring most aggressively for AI-related roles are healthcare, finance, law, and manufacturing. Healthcare is hiring because AI can help diagnose diseases, predict patient outcomes, and manage records — but hospitals need people who understand both medicine and AI to implement it safely. Finance is hiring because AI can detect fraud, manage portfolios, and assess credit risk — but banks need people who understand both finance and regulation to deploy it correctly. Law firms are hiring because AI can review contracts and research cases — but lawyers need people who understand both law and AI to use it without liability.

Manufacturing is hiring because AI can optimize production, predict equipment failures, and manage supply chains. But factories need people who understand both manufacturing and AI to set it up and troubleshoot it. These roles often pay more than the national average because they require both technical knowledge and industry informed.

What skills matter for the new jobs

You do not need a computer science degree to move into AI-related work. Data annotators need attention to detail and the ability to follow instructions — many companies train people on the job. Prompt engineers need curiosity and the willingness to experiment; many learned the skill by playing with AI tools on their own. AI trainers need domain knowledge in their field plus the ability to think critically about whether an AI system is working correctly.

What matters more than credentials is the ability to learn. AI is changing fast, and someone hired as a prompt engineer today might be doing something different in two years. Companies are looking for people who can adapt, ask good questions, and pick up new tools quickly. Many of these roles did not exist five years ago, so there is no traditional career path. You learn by doing.

What this means for your job search

If you work in customer service, data entry, or routine writing, it is worth thinking about what comes next. That does not mean your job will disappear tomorrow — many of these transitions happen over years. But it means the market for those roles is shrinking, and wages may not grow as fast as they used to. Learning a skill that complements AI — like evaluating whether AI output is correct, or managing AI systems in your industry — makes you more valuable.

If you are starting out, look at where growth is happening. Healthcare, finance, and manufacturing are all hiring for AI-adjacent roles. You do not need to become a software engineer. You could become a data annotator, move into training, and eventually move into management. Or you could stay in your current field — nursing, accounting, manufacturing — and learn how AI works in that field. That combination of domain knowledge plus AI literacy is increasingly valuable.

Frequently Asked Questions

Will AI take my job?

It depends on what you do. If your job involves routine, repetitive tasks — answering the same questions, entering data, writing basic copy — the risk is higher. If your job requires judgment, complex problem-solving, or deep knowledge of your field, the risk is lower. Most jobs are changing rather than disappearing entirely, which means the role is evolving but positions still exist.

What is the fastest way to move into an AI-related job?

Data annotation is the easiest entry point because it requires no prior experience and many companies hire remotely. You can start there, learn how AI systems work, and move into training or evaluation roles. Alternatively, if you already work in healthcare, finance, or law, learning how AI is used in your field makes you valuable without requiring a career change.

Do I need a degree to get hired for these roles?

No. Data annotators, prompt engineers, and many AI trainers do not need degrees. What matters is that you can demonstrate the skill — either through a portfolio of work or by passing a test. Some companies care more about what you can do than what credentials you have.

Are AI jobs paying well?

It varies. Data annotation pays $15 to $25 per hour, which is modest. Prompt engineering and AI training pay $40,000 to $90,000 per year depending on experience and company. Compliance and ethics roles in regulated industries pay $70,000 to $130,000 per year. The highest-paying roles usually require both AI knowledge and informed in another field.

What should I learn right now if I want to work with AI?

Start by using AI tools yourself — ChatGPT, Claude, or similar systems. Understand what they can and cannot do. If you work in a specific field, learn how AI is being used in that field. If you are starting from scratch, consider learning basic data analysis or how to evaluate whether AI output is correct. Most of these skills you can learn through free online courses or by experimenting on your own.