Generative AI models have recommended insecure source code for 45% of tasks.
According to Veracode, Java code failed security checks the most, with a 72% failure rate.
The Veracode study analyzed outputs from more than 100 LLMs released over the past three years.
We Asked 100+ AI Models to Write Code. Here’s How Many Failed Security Tests. | Veracode
How secure is code generated by AI? We asked 100+ AI models to write code. Here’s key findings from the 2025 GenAI Code Security Report.
Veracode (www.veracode.com)
