Global Coding Surge Linked to AI Slop Influx as Security Vulnerabilities Reach Critical Threshold

AI tools allow anyone to code, but "vibe coding" is fueling a tsunami of "slop" and security risks. Experts warn of a 70% increase in critical software errors.

By: AXL Media

Published: Apr 8, 2026, 4:45 AM EDT

Source: Information for this report was sourced from NBC News

The Rise of Vibe Coding and Developer Throughput

Artificial intelligence has fundamentally altered the software engineering landscape, enabling users with zero computer science experience to launch functional apps and websites. This phenomenon, often referred to as vibe coding, prioritizes developer productivity and the rapid shipping of new features over traditional manual labor. Within leading tech firms, the shift is even more pronounced, with some senior engineers reporting that AI now generates 100 percent of their code. According to Boris Cherny of Anthropic, human developers are increasingly acting as high level architects rather than manual mechanics, though this transition brings unique structural risks.

The Proliferation of AI Generated Slop

As the volume of automated code increases, the industry is facing a wave of low quality output known as AI slop. Unlike human programmers who may make simple typos, AI systems tend to struggle with the broader maintainability and readability of large repositories. David Loker of CodeRabbit noted that AI often fails to recognize existing functions within a codebase, leading to the duplication of logic in multiple locations. This sprawling complexity creates a significant burden for maintenance, as updates to a single business rule must be manually tracked across various redundant sections to prevent logic conflicts.

Security Risks in the Era of Automated Programming

The convenience of AI coding agents is frequently traded for essential security standards, according to recent research. Experts suggest that while a large language model may be proficient at writing individual lines of code, the sheer volume of output creates an unmanageable attack surface. Jack Cable of Corridor emphasized that even if AI writes code more efficiently than a human, the 20 fold increase in production requires a commensurate increase in rigorous review. Without expert oversight, coding novices are gaining significant public exposure without the necessary security expertise to protect user data from sophisticated threats.