According to a recent TechRadar report summarizing Salt Security research, nearly all security executives express concerns about the safety of AI-generated code. Despite this apprehension, about a third of organizations still rely mainly on manual code reviews before deployment, highlighting a governance gap as AI significantly speeds up software production.
- 90% of security leaders concerned about AI code risks
- 29% identify insecure coding patterns as primary AI threat
- Over one-third still depend on manual code reviews
Product angle
The TechRadar source highlights that AI coding assistants have been rapidly adopted across development teams, outpacing existing security frameworks. This rapid adoption introduces novel risks, as AI-generated code can quietly replicate security flaws embedded in its massive training data. According to the report, current security practices, especially manual code review, struggle to provide effective oversight on AI output, which is produced at volumes too high for human teams to inspect thoroughly.
Security leaders interviewed for this report express growing awareness that traditional governance models are insufficient for AI-driven development environments. The reliance on manual review, although prevalent, is increasingly seen as unsustainable and inconsistent. Experts suggest that rethinking AI coding tools as supply chain components subject to rigorous vetting might offer a necessary path forward for better managing these emerging risks.
Best for / avoid if
Organizations investing in accelerating their software delivery through AI coding assistants may find value in awareness of the rising security concerns detailed in the source report. Teams looking to innovate quickly yet maintain security rigor should consider supplementing or replacing manual review with automated and continuous security validation tools tailored to AI-generated code. Larger enterprises facing challenges in managing distributed workflows might particularly benefit from modern governance frameworks aligned with AI's impact.
Conversely, companies that currently rely solely on manual code auditing before launch may struggle to scale securely with increased AI-assisted development. Firms lacking resources to adopt integrated security automation or update governance processes could expose themselves to heightened vulnerability risks. Organizations with strict compliance requirements or those in highly regulated industries should be cautious about uncritically adopting AI code generation without enhanced security controls.
Pricing and alternatives to check
While the TechRadar summary does not specify pricing details, the discussion implies the need for investment in advanced security tools that complement AI development efforts. Prospective buyers should evaluate solutions that integrate continuous code analysis, AI-specific vulnerability detection, and automated governance workflows to build confidence in AI-generated software pipelines.
Alternatives to purely manual code review include security platforms offering AI-driven code scanning and threat modeling capabilities. Vendors specializing in software supply chain security or DevSecOps automation may provide better scalability and consistency. Buyers are advised to compare these offerings against their current manual processes and consider industry benchmarks reflecting AI risk management best practices.