New Techreviewer.co Research: 89% of Software Companies Now Use AI to Write Code, but 90% Report at Least One Downside
New Techreviewer.co Research: 89% of Software Companies Now Use AI to Write Code, but 90% Report at Least One Downside
The third annual study of 127 software development firms reports record productivity gains, but also highlights increasing concerns about hallucinated code, security risks, and skill erosion among junior developers.
MANCHESTER, Conn.--(BUSINESS WIRE)--Techreviewer.co, a B2B research and ratings platform for technology service providers, has released its third annual research on AI use in software development, polling 127 companies worldwide.
97% of software companies report productivity gains from AI, but 90% have also hit at least one downside, from hallucinated code to security vulnerabilities.
Share
The study shows that 89% of software development companies now use AI to write or assist with code, with the median company reporting that AI manages 26% to 50% of its codebase. About one in four companies say AI produces more than half of their code, and 62.2% report daily AI use by most or all developers. On average, companies use four AI tools simultaneously. Claude/Claude Code leads adoption at 93.7%, followed by ChatGPT/OpenAI tools (77.2%), Google Gemini (58.3%), GitHub Copilot (57.5%), and Cursor (52.0%).
Nearly 97% of respondents reported productivity gains from AI, with no company reporting a decline in any of the past three years. The proportion of companies reporting gains above 50% has increased from 7.5% in 2024 to 30.7% in 2026.
However, about 90% of companies reported at least one downside. More than half (52.8%) cited hallucinated or incorrect AI suggestions, 44.1% noted increased code review workload, and 33.1% experienced security or vulnerability issues in AI-generated code. Over a third (37.0%) identified over-reliance on AI or skill erosion, especially among junior developers, which is now a more common concern than technical debt (23.6%) or increased defects (22.0%). Data privacy and security were the top implementation challenges, cited by 58.3% of respondents, while 15.0% reported having no formal AI training strategy.
Companies are increasingly building AI expertise internally. In-house training rose from 57.5% in 2024 to 72.4% in 2026, while hiring AI specialists declined from 35.0% to 15.0%. Nearly all respondents (97.7%) expect AI's role in development to continue growing over the next five years.
The full report, methodology and expert commentary are available at: https://techreviewer.co/research/ai-in-software-development-in-2026
About Techreviewer.co
Techreviewer.co is a B2B research and ratings platform that helps businesses evaluate and connect with software development, IT services, and AI technology providers.
Contacts
Media Contacts
Yury Zelianko, Product Manager
y.zelianko@techreviewer.co
