Evidence-based research and analysis from C3S Consulting.
An evidence-based assessment of the near-, medium-, and long-term costs of prioritizing public, general-purpose AI over human expertise — in dollars, security exposure, code quality, and human capital — with a survey of the private and specialized alternatives that recover that value.
Key findings: 95% of GenAI pilots showed no measurable P&L impact (MIT NANDA). Shadow AI added ~$670K to average breach cost (IBM). 45% of AI-generated code introduced OWASP Top-10 vulnerabilities (Veracode). AI made experienced developers 19% slower while they believed they were 20% faster (METR).
Sign up for the C3S newsletter and read the full paper immediately.