AI Code Review Bottlenecks: Can Engineers Trust AI-Generated Code?

AI-generated code review has become a significant challenge for software engineers, with 96% of developers not fully trusting its output to work correctly.

The Rise of AI-Generated Code

AI coding tools can now generate thousands of lines of code in minutes, helping companies build features, run tests, and fix issues faster. However, this rapid code generation has created a new bottleneck: reviewing the AI-generated code. According to a survey of over 1,100 developers by Sonar, an AI code-verification startup, respondents estimated that AI contributed 42% of the code they added to shared codebases.

The Problem with AI Code Slop

While AI-generated code can look clean on the surface, it often conceals sloppy mistakes such as faulty assumptions, security vulnerabilities, or subtle errors that emerge only after deployment. Fixing these problems could erase the productivity gains AI promises. Companies are responding to the onslaught of AI code slop by rethinking how they review code.

Strategies for AI Code Review

New strategies are emerging to address the AI code review challenge. Among other approaches, engineers are:

1. Scrutinizing plans before AI begins coding: This involves reviewing the requirements and specifications before handing them over to AI tools to ensure that the code generated meets the necessary standards.
2. Deploying specialized AI agents to catch routine flaws: These agents can help identify common errors and inconsistencies in the code, reducing the workload for human reviewers.
3. Sending risky changes to human reviewers: This approach involves flagging changes that are deemed high-risk and sending them to human reviewers for additional scrutiny.
4. Requiring developers to defend the code their agents produce: This involves having developers explain and justify the code generated by AI tools, ensuring that they understand the underlying logic and can defend it.

The Future of Code Review

The new era of code review will determine whether AI can ever provide code that is both faster and more reliable. It also has some software engineers thinking about the future of their profession: If entry-level engineers spend less time writing code themselves, how will they learn to judge it?

AI Code Review Bottlenecks

AI-written code is shifting the bottleneck from generating software to reviewing it. According to the Sonar study, 38% of developers said “more effort” is required to review AI-generated code than code written by their colleagues. Sixty-one percent of them said AI often produced code that looked correct but was “unreliable.”

The Case of Synthesia

For Synthesia, an AI video-generation platform, code review has become essential to its engineering workflow. In November 2025, Synthesia’s 118 engineers went all-in on AI coding tools like Claude Code. According to Peter Hill, Synthesia’s chief technology officer, the result has been a massive surge in code volume.

Duplication and Redundancy

One recurring problem is duplication. Hill says AI tools may not recognize that code for a task already exists, and they’ll write another version because they have limited context. Synthesia has found as many as 10 versions of the same function, leaving engineers to identify and remove redundant functions.

The rise of AI-generated code has created a new bottleneck in the software development process: code review. While AI tools can generate code quickly and efficiently, they often produce code that is sloppy, insecure, or unreliable. Companies are responding to this challenge by rethinking how they review code and implementing new strategies to address the AI code review bottleneck.