Four in 10 developers have shipped AI-written code that they did not fully understand, according to a survey of 1,000 developers by software development and AI consultancy Sombra.
Sombra uses the term ghostcode for code that reaches production without any human understanding it in full. Its figures suggest the review step is weaker than it looks. 58% of respondents said they would need AI's help to find a bug in their own code, and 52% said the same about reviewing someone else's.
Confidence in working without the tools runs ahead of ability. 73% said they would be fine without AI tools, yet 45% said they would need AI to explain how code they wrote themselves works.
Yurii Nakonechnyi, Sombra's CTO, compared the position with 2010, when open source application frameworks spread and developers used them without knowing how they worked underneath, which made debugging and review harder. The difference now, he said, is that the codebase itself becomes the framework.
To stay across the latest in cloud, AI and enterprise tech analysis from Compare the Cloud, subscribe to our weekly newsletter at https://www.comparethecloud.net/newsletter
