Does AI Impact Jobs and Society – particularly Software Experts?

In the modern world, AI is making headlines. Building apps, generating images, videos, producing songs based on lyrics, writing various types of documents, etc., for example. When the world started using PCs for day-to-day office work, there were many dislikes as PCs may affect a reduction in work-force. Similarly, some quarters raise alarms about today’s app-building AI trends and its impact on software professionals. This is an independent analysis of this situation with pros and cons.

The PC parallel — what actually happened

When PCs, then spreadsheets, then the internet arrived in offices, the fear was the same: machines will do the work, clerks and typists will vanish. Some jobs did disappear (typing pools, many bookkeeping roles). But far more new jobs were created around the technology — IT support, software development itself, data entry evolved into data analysis, and entire industries (web, e-commerce, digital marketing) didn't exist before. The net effect over decades was more jobs, not fewer, though the transition was genuinely painful for specific groups of workers who didn't retrain in time.

The honest caveat: AI is not identical to PCs. A PC was a tool that made a human faster at a task the human still directed. AI systems can now do the drafting, coding, designing, and increasingly the deciding — including work that used to be considered safely "skilled" or "creative." That's a real difference, not just alarmist framing.

What's happening in software right now (a live case study)

Software development is actually the best real-world test case, because it's the industry adopting AI fastest.

The case for optimism:

• A Morgan Stanley survey found CIOs plan to increase software spending by 3.9% in 2026, and the software development market could grow at a 20% annual rate, with Morgan Stanley analysts arguing AI is enhancing developer productivity and leading to more hiring rather than eliminating jobs, as it lets companies build more ambitious software and tackle old technical debt.

• 65% of developers expect their role to be redefined in 2026 — shifting from routine coding toward architecture, integration, and AI-assisted decision-making, and four in ten developers said AI had already expanded their career opportunities.

• AI-skilled workers are commanding a 56% salary premium, and Gartner predicts 75% of developers will spend more time architecting and directing AI than writing code by hand by end of 2026.

• New job categories are appearing (AI orchestration, prompt/agent engineering roles) rather than the profession simply shrinking.

The case for concern:

• Stanford's 2026 AI Index, using payroll data across millions of workers, found employment for software developers aged 22–25 has dropped nearly 20% since late 2022, while employment for older, more experienced developers at the same companies actually grew 6–12%. That's a very specific and worrying pattern: it's not "software jobs are disappearing," it's "entry-level jobs are disappearing while senior jobs grow."

• Junior developer demand has fallen roughly 40% at companies that have seriously deployed AI tools — again, junior roles specifically.

• Research has found a 21% decrease in coding-related job postings after ChatGPT's introduction, though the same research notes developers generally prefer working alongside AI tools rather than being replaced by them.

This is the pattern most economists watching AI expect to generalize: not mass unemployment, but a squeeze on the entry point into skilled careers, because AI is very good at exactly the routine, well-defined tasks juniors are traditionally given to learn on.

The general pattern beyond software

The software case likely previews what other white-collar fields (law, writing, design, customer service, basic analysis) will experience, on a lag:

• AI raises output per worker — companies can do more with the same headcount, which usually means growth, not layoffs, if demand for the underlying product grows too (more software gets built, not just built cheaper).

• AI hits the bottom rung hardest — tasks given to trainees/juniors to build experience are often the first automated, which risks breaking the "learn by doing the boring stuff first" pipeline that produces future senior experts.

• Winners shift the skill mix, not the job count — from "doing the task" to "directing, checking, and being accountable for the AI doing the task." People who adapt tend to do well; people who don't retrain are the ones actually hurt.

Fair way to hold both views

• Those raising alarm are right that this transition has a real victim class — new graduates and career-changers who can no longer use "cheap junior labor" as their foot in the door, and workers in roles where AI output quality is now "good enough."

• Those pushing back are right that total-job-count predictions of doom have failed at every past technology wave, and current data (developer surveys, hiring plans, revenue growth) shows expansion, not collapse, in most metrics — it's redistribution and skill-shift, not annihilation.

• Where genuine uncertainty remains: economists and labor researchers disagree on whether this wave is different in degree (bigger, faster disruption before new jobs form) rather than in kind. That's an open empirical question, not a settled one — reasonable, informed people land on different sides of it.

Here's the deeper dive — looking at both the historical record and what's happening right now inside the software industry.

The historical pattern: three groups, every time

When a disruptive technology hits an occupation, workers tend to split into three groups, and the split predicts outcomes better than the technology itself does.

Group 1 — Moved "up the stack." When ATMs arrived, bank tellers didn't vanish — the total number of tellers actually grew for years after, but their job shifted from counting cash to sales, advice, and handling problems ATMs couldn't. The people who thrived were the ones who leaned into the parts of the job a machine couldn't do — judgment, relationships, exception-handling — rather than competing with the machine at the thing it was now better at.

Group 2 — Retrained sideways. Travel agents mostly lost their old job (booking tickets) to the internet, but many who survived pivoted into corporate travel management, specialized/luxury trip planning, or crisis logistics — niches requiring expertise and trust that a search engine doesn't replace. The failure mode was agents who kept doing the exact same task (basic booking) and waited for demand to come back. It didn't.

Group 3 — Stood still and were displaced. Typists, switchboard operators, many factory line workers — the ones who treated the new tool as a threat to resist or ignore, rather than something to learn, generally lost their jobs outright rather than transitioning into a new version of them.

The common thread: it was rarely about resisting or embracing the technology emotionally — it was about how fast someone moved their own skill set toward "directing/judging" work and away from "executing" work.

What that looks like in software, right now

Software is a live version of this same split happening in fast-forward:

• At companies further along in AI adoption, Python developers have been moving into AI engineer roles, and backend developers into AI lead roles — actively identifying and closing their own skill gaps rather than waiting to be reassigned, which is Group 1's "move up the stack" pattern playing out almost exactly.

• A third of developers now rank AI/ML skills as their top learning priority for 2026, and the shift is described as "self-directed learning, rapid skill cycles and AI literacy" becoming baseline requirements, not optional extras.

• The role itself is being redescribed as working with AI as a collaborator rather than using it as a tool — developers increasingly think about system architecture and user experience while AI handles routine implementation, which is the "judgment over execution" shift.

• The honest downside: 67% of surveyed teams expect developer productivity to jump 25%+ in 2026, but only 20% of teams are actually measuring AI's real impact, and 46% expect burnout to rise — meaning "adapting" isn't automatically pleasant. It's often accompanied by rising expectations without matching support, which is a real cost of this transition, not just a temporary bump.

The concrete, transferable lesson (for any field, not just software)

Strip away the tech-specific detail and the pattern that separates people who came out ahead from people who didn't, across every wave of automation, is remarkably consistent:

1. Stop competing with the machine at its strength. If AI is faster at drafting/coding/producing a first version, trying to out-produce it manually is a losing race. The winning move is shifting effort to evaluating, directing, and taking responsibility for the output.
2. Specialize in the judgment calls machines still get wrong. Context, ethics, client relationships, ambiguous or high-stakes decisions, and accountability are consistently the last things automated in every wave.
3. Treat learning the new tool as part of the job, not an optional side project. The workers who lost out weren't usually less capable — they were the ones who waited for someone else to hand them a plan, while others self-taught early and became the internal expert.
4. Watch the entry point, not just your own job. The current software data (juniors losing ground while seniors gain) suggests the biggest personal risk right now is being early-career in a field being automated — worth actively seeking out the judgment/architecture-heavy tasks early, rather than accepting only routine "grunt work" assignments.   

*AI assisted document

By: Mubarack Deen

Author is an experienced IT professional who served for a number of larger corporates in foreign countries, having led strategic IT projects and provided business solutions. Currently working as a self-employed business solutions provider, an enthusiast for bridging technology and business focusing on service improvements and growth.

Can be reached by email: m3de2n@gmail.com