AI-Generated Code Overload: How Corporations Are Struggling with the Surge in Automated Development

Corporations are rapidly embracing AI to churn out vast amounts of code, positioning it as a revolution in productivity. Yet, a behind-the-scenes investigation by The New York Times reveals a far less glamorous reality. Programmers are drowning in AI-generated code they can’t keep up with, while employers scramble to find ways to verify the hasty output.

One financial services company, for example, saw its coding output skyrocket tenfold after adopting the popular AI tool Cursor. The result? A backlog of one million lines of code requiring review, according to Joni Klippert, CEO of the security startup StackHawk, which collaborates with the firm. Ignoring this glut is not an option. Poor-quality code—whether AI-generated or human-written—can cripple software and introduce security vulnerabilities.

Recent incidents at Amazon and Meta highlight the risks. Both companies experienced disruptions after AI tools took unauthorized actions, and these are just the cases that have become public. “The sheer amount of code being delivered, and the increase in vulnerabilities, is something they can’t keep up with,” Klippert told The New York Times. The surge in AI-generated code has also spilled over into other departments, such as sales and marketing support, creating “a lot of stress”.

The Paradox of AI in the Workplace

We’ve reached a critical inflection point in AI’s impact on the workplace. Companies have used AI to justify workforce reductions, with reports indicating that AI was cited in the announcements of over 54,000 layoffs last year. This year alone, major tech firms like Block (founded by Jack Dorsey) and Atlassian laid off thousands of employees while emphasizing AI-driven pivots.

Ironically, while jobs are being eliminated, AI is simultaneously creating new demands for human oversight. Someone must test the AI-generated code, a task traditionally handled by the original developer. However, programmers are now too occupied with prompting AI tools to take on this additional responsibility. The question remains: Who will fill this gap?

“There are not enough application security engineers on the planet to satisfy what just American companies need.” — Joe Sullivan, adviser to Costanoa Ventures

Worse still, AI may be making programmers’ jobs harder. Software engineers report that the pressure to produce more code while constantly supervising AI tools is accelerating burnout—a phenomenon researchers have dubbed “AI brain fry.”

Companies Scramble to Address the Code Glut

Businesses are still figuring out how to manage the flood of AI-generated code. Michele Catasta, president and head of AI at the startup Replit, summed up the dilemma: “The blessing and the curse is that now everyone inside your company becomes a coder.”

Sachin Kamdar, CEO of the AI agent startup Elvix, advocates for a strict human review process, stating that fixing errors later would be far more difficult.