Software engineers across the US and Europe have begun treating artificial intelligence tools as theater. Half of them turn to these systems mainly to signal compliance to management. They worry that skipping the performance could cost them their positions.

This pattern emerged clearly in fresh research released today. A survey of 1,000 developers in five countries laid bare the gap between corporate ambition and daily reality. Organizations push adoption targets. Engineers respond with visible but often unproductive activity. TechRadar first detailed the findings from The Adaptavist Group.

Fifty percent admitted they use AI primarily to reassure bosses and project an image of innovation. Forty-seven percent reported their performance reviews now hinge more on frequency of AI engagement than on code quality. And 58 percent said AI usage has entered their formal KPIs. The numbers paint a picture of compliance over competence.

But the pressure runs deeper. Fifty-four percent expressed heightened concern about long-term career prospects. Thirty-seven percent voiced outright fear of job loss if they fail to demonstrate AI proficiency. Seniority offered no shield. The anxiety held steady from early-career to experienced developers.

Sixty percent felt pushed to achieve more with fewer resources because of these expectations. For those under forced adoption the figures climbed. Sixty-one percent reported greater career anxiety. Thirty-six percent said the stress now harms their actual output. The data comes from the same Adaptavist study, also covered in an EIN Presswire release.

Matt Saunders, DevOps lead at Adaptavist, captured the mismatch. “A good developer experience matters for psychological well-being and for organizational growth and innovation,” he noted. Companies appear to have rolled out AI without enough regard for existing workflows or human impact.

The result looks like an impending talent flight. Developers describe eroded engagement. Some speak of using tools more often than needed simply to meet visible quotas. The behavior satisfies metrics. It rarely advances meaningful progress.

Separate research released this week reinforces the unease. Stack Overflow’s 2026 Developer Survey polled more than 30,000 professionals across 169 countries. Nearly half described themselves as complacent in their roles. Another third said they felt unhappy. Only 22 percent reported genuine happiness at work.

Burnout and technology fatigue topped reasons for complacency, cited by 21 percent. Economic uncertainty followed closely. AI automation and stalled advancement also weighed on respondents. Business Insider examined the survey’s implications just yesterday.

Adoption itself runs high. Almost two-thirds of developers now use AI coding assistants or agents. Sixty-two percent rely on general-purpose chat tools. Thirty-one percent of daily users spend four hours or more with them. Yet trust remains limited.

Fewer than 10 percent would trust AI for important decisions. Nearly 80 percent stressed the need for clear source attribution on generated code. They value verification. Many spend increasing time reviewing output they did not create.

Marc Backes, senior software engineer at Directus, described the shift. He once found satisfaction in tracing problems to their source or building features from scratch. Now much of that work goes to agents. “It’s definitely less fulfilling,” he told Business Insider. The pressure to keep pace feels industry-wide rather than managerial.

Executives tell a different story. Coinbase CEO Brian Armstrong admitted in 2025 that he fired engineers who resisted signing up for the company’s AI coding tools. He issued a direct Slack mandate and followed through with accountability meetings. GitHub CEO Thomas Dohmke has been equally direct. “Either you embrace AI, or get out of this career,” he posted on X while sharing related research.

Some companies report dramatic productivity gains. Others cite AI when announcing staff reductions. Block cut thousands and pointed to new ways of working enabled by the technology. Similar language appeared at Atlassian, Meta and Amazon. Yet broader labor data shows software developer employment still growing, even if at a slower rate than pre-2022 trends.

Analysts debate whether current layoffs represent genuine displacement or “AI washing” of financial decisions. A Federal Reserve review and research from Anthropic found limited evidence of widespread job losses tied directly to AI so far. Productivity in software development has accelerated since ChatGPT’s arrival. The developer population worldwide has expanded by 20 to 50 percent depending on the estimate.

The human cost appears real nonetheless. Engineers report identity struggles. They move from creators to reviewers and orchestrators. Some describe themselves as “human meat proxies” who press buttons on AI-generated plans, tests and code. The work can feel soul-crushing after years of deep technical problem solving.

Others have begun to push back. A growing minority publicly reject AI tools even when it risks clients, offers or employment. Brett Chalupa posted a viral video declaring he would no longer code with AI after 20 years in the field. He cited loss of craft, purpose and broader societal concerns. Several developers have left roles or the industry altogether rather than adopt the new workflow.

Yet the majority feel they cannot afford to opt out. The message from leadership is clear. AI is not optional. Performance systems now track it. Budgets favor it. And the pace of change shows no sign of slowing.

Organizations face a choice. They can continue tying compensation and security to visible AI activity. Or they can refocus on outcomes, code quality and sustainable developer experience. The Adaptavist data suggests the current path risks disengagement, higher turnover and technical debt that surfaces later.

Productivity metrics have improved in some settings. Review cycles now consume more time than before, sometimes erasing initial speed gains. Bain & Co. noted developers complete 21 percent more tasks with AI but spend 91 percent more time reviewing. The bottleneck simply moved downstream.

Developers who integrate tools thoughtfully report higher ambition and creativity once past initial skepticism. Those forced into superficial use feel the opposite. They sense skills eroding. They question their future value. And they watch as senior voices warn that resistance equals obsolescence.

The tension will not resolve quickly. New surveys arrive almost weekly showing the same pattern: widespread adoption paired with deep reservations. Burnout climbs. Happiness falls. Anxiety spreads across experience levels.

Companies that treat AI as a pure efficiency lever may lose the very talent needed to guide it. Those that invest in training, realistic expectations and measurement tied to business results stand a better chance of retaining engineers who still want to build things that matter. The data is in. The question now is whether leadership will read it.