Job postings for entry-level software developers declined as much as 73% from February of 2022 until February of this year, while postings for more specialized AI roles increased by 156% over the past year, with median salaries around $177,000.
This article provides a methodical investigation of task-based automation statistics, job creation patterns, and wage information to help establish the actual job security of technology jobs and answer exactly what tech jobs are safe from AI.
Why Entry-Level Tech Jobs Are Disappearing
The path to entry into the technology industry has closed off for entry-level workers rapidly. Historically, entry-level opportunities in tech have consisted of basic digital tasks such as writing CRUD-style applications, monitoring system alerts, and collecting routine data from market research.
Those same types of activities are now the types of tasks being targeted most heavily by large language models. Since early 2023, posted job listings for entry-level tech positions have decreased about 35%. Compared to 2019, the number of new graduates that tech companies bring on board each year has decreased over 50%, and by 2024, new graduates to total new hires will be approximately 7%.
This significant decline in entry-level hiring has created a significant structural problem in the overall job market. Employment rates for junior workers in tech at age 22 to 25 are now approximately 19% lower than what they would generally be expected to be based on average national job creation rates of 2.5%.
Additionally, while the overall national average for job creation is approximately 2.5%, the job-finding rate for entry-level workers in the most exposed areas of technology jobs declined by approximately 14% when compared with 2022 levels. According to GitHub's AI Job Risk Database, junior software developers now have an estimated 60% probability of being completely automated within three to five years.
This transition has not gone unnoticed by the executive community. 63% of business leaders anticipate that AI will completely automate repetitive entry-level functions over the next 3-4 years. Consequently, the foundation of the tech career ladder has dropped out. It will now be extremely challenging to gain entry into this workforce.
The Difference Between Automating Tasks and Replacing Jobs
AI will automate only certain functions associated with specific job types instead of replacing entire categories of jobs. Many market analyses of tech job types are erroneous because they treat a job type as though it is a single piece of work with no interdependent components. Utilizing the O*NET task mapping schema to map out job types reveals a very different picture.
For example, there are many computer programmer job types, with 74.5% of this job's tasks automated. Algorithms such as GitHub Copilot, Cursor, and Claude Code now provide programmers with capabilities to write boilerplate code, set up CI/CD pipelines, and produce testing scripts.
The Bureau of Labor Statistics predicts that the number of computer programming jobs will decrease by 6% until 2034. However, just because a significant portion of a computer programmer's work will be automated does not mean that engineering departments will disappear. Instead, they will transform from writing raw code daily to evaluating the outputs of machine learning models, integrating varied systems, and addressing technical debt.
Ultimately, the "true measure" of job security will be based on the demand for human responsibility and ownership for a company's computer systems. Algorithms may generate Terraform scripts for cloud provisioning, for example. However, determining a rollback strategy during a database disaster will be subject to the judgment, ethics, and accountability of a human being. Algorithms cannot be held accountable for server downtime.
They will not be able to negotiate with upset business users regarding system downtime.
What Tech Jobs Are Safe from AI: Risk and Growth Explained
In the world of work, there exists an unusual paradox of some jobs having a high level of task exposure risk but a very strong growth in their market values, as evidenced within the DevOps engineering job description.
Many risk databases have scored the DevOps engineering job descriptions anywhere from 40% to 66% task exposure. Much of their daily work consists of routine pipeline maintenance, cloud cost reporting, and configuration tasks, which can easily be completed through the use of automation tools.
Even with this level of exposure for DevOps and cloud engineering jobs, the projected growth for the DevOps and cloud engineering job categories is 18% over the next 10 years. This is due to the simple fact that a company's overall cloud architecture and the reliability of the systems they build continue to be crucial to the overall success of a business.
The role of DevOps and cloud engineers is evolving from a manual configuration perspective to that of managing complex automated deployment engines.
Technology jobs that require high levels of tolerance to ambiguity, cross-team collaboration, and strategic decision making fall within the safe zone. These types of jobs are those that have a very high cost associated with a mistake being made. When wondering what tech jobs are safe from AI, look for roles that carry heavy consequences for failure.
Cybersecurity and Incident Response
Cybersecurity governance and incident response have been structurally very safe due to the fact that the cybersecurity industry has experienced a tremendous increase in job openings and employment opportunities. Of the approximate 601,000 U.S. workers within the cybersecurity field, about 70% are currently sitting within the green/safe zone at an average domain risk score of 49 out of 100. However, career safety within cybersecurity is highly contingent on the specific sub-role.
As the demand for Tier-1 Security Operations Centers (SOC) continues to decline rapidly, the majority of security professionals (56%) indicated that the introduction of automation has reduced the need for entry-level staff to perform basic alert triage functions. In contrast, senior positions such as security architect, threat hunter, and incident commander remain relatively protected.
A human incident commander must always make the final decision whether or not to disconnect a live hospital database during a ransomware attack even if an anomaly has been flagged by a network model.
Cloud Infrastructure and System Reliability
Site Reliability Engineers (SRE) and cloud architects will take care of your entire company's digital system's health and reliability. Basic server provisioning is currently automated.
However, SRE teams will evaluate legacy on-premise systems and design a modern cloud-based solution through strategic trade-off analyses that cannot be computed by models. A good system design requires knowledge of your company's budget, employee skill sets, and plans for expansion in the future.
User Experience Design and Product Management
There is much ambiguity in product management and User Experience (UX) design. Clients will typically give vague requests, and the product manager and UX designer will interpret those requests into usable engineering requirements. In addition to the vagueness of the client's requests, product managers and UX designers have to balance their client's business objectives with ethical limitations, market trends, and the customer's unpredictable behavior.
Although models can provide quick wireframe prototypes and summarize the results of market research, they cannot negotiate with human stakeholders on a product roadmap nor identify the key emotional friction points in a user's journey.
Technology Risks and the Cost of Automation
In order to comprehend technological risk, you have to consider the economics of automation. Anthropic claims that physical robots have the capabilities to perform three-quarters of all physical job tasks in the United States. But currently, they only have the economic competitiveness for 0.3% of those tasks, and the requirements to manufacture, ship, and repair the physical hardware are extremely costly.
Thus, the most highly paid jobs with the least risk of AI displacement, based on the market trackers, is the electrical power-line installer with a median salary of $92,560. The nurse anesthetist is also the highest rated AI-Resistant Index score of 93.3 and median salary of $195,263.
In the digital realm, the economics are completely reversed. The marginal cost of duplicating software is extremely close to zero.
An AI model would be able to efficiently analyze millions of lines of code, for example, and a sophisticated AI could analyze several hundred million lines of market data at a price of less than 1 cent per trade. For this reason, digital positions whose primary function is to process data, including both data processing (with average salaries of $117,000) and basic financial securities analysis (with average salaries of $109,000), represent the strongest potential for automation.
Remote Work Options for Offshore Teams
The rapid changes to automation create specific challenges and opportunities for technology professionals based in Southeast Asia, particularly those in the Philippines. The challenge for offshore workers in the Philippines, therefore, is that they cannot simply target jobs related to electricity, such as power line repair, as they cannot be automated by AI.
As such, the strategy to tap into offshore workers must focus on finding remote-friendly, high-safety digital jobs. Cloud security, SRE, and product management jobs currently are the primary areas of demand for offshore workers serving US and EU employers. Remote workers currently working in either entry-level data entry or Tier-1 IT support roles will not have access to cloud cost reporting, zero trust security auditing, or identity access management, and therefore will not retain any global market value.
To figure out what tech jobs are safe from AI, offshore workers must focus on these complex cloud and security roles.
Training Paths for At-Risk Careers
The hope that everything will work out for the best is not a valid plan for creating a career. Workers in very high-risk and vulnerable positions need to have a career pathway mapped. The transition from being a vulnerable junior-level worker to being a secure senior-level employee takes careful thought regarding the development timeline required and the skills required to fill those desired roles.
The following is a 12-to-18-month trajectory for four of the most common tech job roles that have the potential to be eliminated:
SOC Analyst Level 1 to Security Architect: Transitioning from manual alert response to proactive automated threat hunting and creating zero-trust policies using industry-standard certifications, such as SANS SEC595.
Junior CRUD Developer to Cloud Site Reliability Engineer: Moving from basic application feature development to mastering Kubernetes, Infrastructure as Code (IaC) practices, and using advanced monitoring systems.
Data Entry Analyst to AI Model Auditor: Changing from data entry and processing tasks to evaluating large language models for prompt injection vulnerabilities and output bias.
Basic Tech Support to Identity and Access Management Specialist: Moving from password resets to developing complex Identity and Access Management governance structures for enterprise-wide networks.
The Future Economy and Job Forecasts
Global economic data and trends are strongly indicating that the workforce will undergo a significant restructuring over the coming years. This will create many new jobs as well as eliminate many others. A report by Goldman Sachs estimates there are potentially 300 million jobs worldwide susceptible to automation through artificial intelligence. According to PwC, an estimated 30% of all occupations could be automated by the mid-2030s.
However, as new technologies emerge, they create new job opportunities, displacing the older occupations. The World Economic Forum estimates that by 2025, 85 million jobs will be displaced but replaced by 97 million new jobs for a net gain of 12 million jobs.
While job postings for software developers decreased significantly over a period of time, they rebounded by 21% shortly after the release of reliable advanced coding tools, like Claude Code, in early 2025. This indicates that companies are still looking to hire qualified engineers for the new AI-generated output.
The key issue is that these newly-created roles are much less evenly distributed than the older roles, with a strong emphasis on experienced, specialized workers.
New AI-model jobs pay a significantly higher salary, with median salaries for AI-related roles estimated at approximately $177,000, compared to an estimated $80,000 for "regular" tech-related roles. Because traditional mentoring through apprenticeships has greatly collapsed, it will be extremely challenging for younger workers to get the experience needed to compete for these higher-paying and safer senior-level roles. When assessing what tech jobs are safe from AI, it's clear that senior, experienced positions hold the advantage.
How to Survive a Career in Technology
Labor market data confirms that the technology sector is diverging into two mutating categories: low-level repetitive work performed with algorithms at a lower cost, and complex tasks requiring the human mind to own the process.
Based on a high probability mathematical risk—as proven by the 73% decrease in job postings for new and/or entry-level developers—the opportunity to have a long-term career in technology will depend much less on avoiding automation than on how one manages automation. Roles requiring strategic decision-making abilities, strict accountability, and substantial cross-team collaboration—such as security architecture, site reliability engineering, and product management—will have substantial value in the job market going forward.
To exist as an employee in this emerging economic environment means to disengage from mechanical output as much as possible, push higher on the complexity envelope, and assume 100% ownership of the outcome of your system.