GitLab has published new findings highlighting what it calls the “AI Paradox,” whereby artificial intelligence is speeding up coding but introducing new productivity barriers as a result. The company’s 2025 Global DevSecOps Report, conducted with The Harris Poll, surveyed 3,266 professionals working in software development, IT operations, and security. The results suggest that while teams deploy faster than ever, they are losing time to inefficiencies that AI alone cannot fix.
According to the study, DevSecOps professionals lose about seven hours per week to inefficient processes. The main causes include fragmented toolchains and collaboration gaps between teams.
Sixty percent of respondents said they use more than five software development tools, and 49 percent reported using more than five AI tools. This tool sprawl, GitLab says, limits the potential gains from AI-driven coding speed.
Eighty-five percent of respondents said that “agentic AI will be most successful when implemented in a platform engineering approach.” GitLab argues that integrated platforms can help align AI with governance and compliance, areas that have become more complex as organizations expand their AI use.
The company notes that 82 percent of organizations now deploy to production at least weekly, amplifying the impact of any friction in development pipelines.
The AI Paradox
The report also suggests that AI is reshaping software roles rather than reducing them. Seventy-six percent of participants believe that as coding becomes easier with AI, there will be more engineers, not fewer. Eighty-seven percent think adopting AI helps future-proof careers, and 83 percent expect their roles to change within five years. Most respondents, a whopping 87 percent, want their employers to invest more in upskilling opportunities.
Adoption is nearly universal, with 97 percent either using or planning to use AI in development. That said, only 37 percent would trust AI to handle daily work without human review. Seventy-three percent have encountered problems with “vibe coding,” where natural language prompts generate code without a clear understanding of its function. Nearly nine in ten agree that creativity and innovation remain human qualities AI cannot replace.
Seventy percent said AI is making compliance management harder, and 76 percent said most compliance issues are discovered after deployment. Forty-three percent view AI for security and compliance as the most important skill for career growth, and 82 percent expect compliance features will be built directly into code by 2027.
“This survey illustrates what we call the ‘AI Paradox,’ where coding is faster than ever, yet the lack of quality, security, and speed across the software lifecycle is causing friction on the road to innovation,” said Manav Khurana, chief product and marketing officer at GitLab. “Toolchain fragmentation has created bottlenecks for developers, and AI agents are amplifying the issue. Organizations need a new framework to match the speed of software development in the age of AI, one that provides intelligent orchestration across the entire software lifecycle while addressing the interconnected requirements of AI orchestration, governance, and compliance that individual point tools simply cannot solve.”
What do you think about GitLab’s findings on the “AI Paradox”? Let us know in the comments.
Image Credit: Yuliia Parovchenko / Dreamstime.com
