AI Code Generation Raises New Risks
Mike Vizard speaks with Mladen Lazic, CTO of Scopic, about AI code quality and the impact of AI-assisted development on software teams. Lazic says AI coding tools can improve productivity, but they also create new trade-offs. Teams are now generating more code at a faster pace. That speed can increase technical debt, code repetition, security gaps and quality issues when the work is not carefully reviewed.
Code Review Becomes the Bottleneck
The discussion highlights a growing challenge for engineering teams. AI can create large amounts of source code quickly, but review and quality assurance still take time. Lazic explains that reviewing AI-generated code can be harder when engineers were not involved in how it was created. Scopic tries to keep developers in the loop from the start. The goal is to guide AI tools like junior developers, not let them build large features without direction.
QA Needs Independent Validation
Lazic says teams should not use the same AI agent to write code, create tests and review the results. That approach can lead to an agent grading its own work. Scopic is moving toward separate agents for coding, pull request review and test creation. Some testing agents do not see the source code. Instead, they work from requirements. That helps validate whether the software meets the original goal, rather than simply matching the code that was written.
Software Engineers Move Toward Oversight
The conversation also looks at the future role of software engineers. Lazic expects AI to remain part of software development. He says engineers will spend more time instructing, monitoring, verifying and improving AI-driven work. That shift makes AI code quality a long-term concern for DevOps teams, security teams and application leaders.

As more of the software development lifecycle becomes agentic, organizations will need better gates, observability and QA practices. Lazic says the goal should not be shipping faster alone. Teams need to ship better software sooner, while reducing technical debt and avoiding hidden security risks.