Home / Technology / AI Security: Leaders Urged to Prioritize Platforms
AI Security: Leaders Urged to Prioritize Platforms
25 May
Summary
- Security cannot be an afterthought as companies adopt AI platforms.
- Shadow AI risks emerge from employee use of unapproved tools.
- New AI defense strategies must match machine speed threats.
Francis de Souza, COO of Google Cloud, stressed that security must be integral to AI adoption, not an add-on. He advised companies to adopt a platform approach, demanding security, governance, and auditability from the start. De Souza warned of 'shadow AI,' where employees use unvetted consumer tools without oversight, and emphasized that AI strategies require concurrent data and security strategies.
De Souza advocated for a consistent security posture across multicloud environments, noting that most companies, even those selecting a single cloud, interact with multiple cloud services. He pointed out the dramatically reduced time between initial breaches and attack progression, from eight hours to 22 seconds, necessitating AI-native, agentic defenses to counter threats at machine speed.
This shift means security is now a board-level and executive team issue. However, a shortage of qualified personnel and the rapid proliferation of AI-introduced vulnerabilities create significant challenges. Recent reports highlight issues with Google Cloud users facing unexpected charges for Gemini API calls after API key scope changes, with delays in key revocation exacerbating security risks.
