latentbrief
Back to news
General1d ago

92% of Companies Hit by AI Security Breaches Lacked Basic Access Controls

The Decoder1 min brief

In brief

  • IBM reveals that 92 percent of businesses affected by AI security incidents had inadequate access controls, with the AI models themselves rarely being the issue.
    • This suggests that human and system oversight rather than inherent flaws in the AI technology are the primary causes of these breaches.
  • The study highlights a critical vulnerability in how organizations manage access to their AI systems.
  • Proper access controls, such as role-based permissions and regular audits, were often missing or insufficient.
    • This lack of basic security measures left systems exposed, leading to incidents like unauthorized data access or model tampering.
  • Moving forward, experts recommend prioritizing robust access control frameworks when implementing AI technologies.
  • Organizations should also consider regular security audits and employee training to mitigate risks.
  • As AI adoption grows, focusing on these foundational security practices will become increasingly essential.

Terms in this brief

access controls
A set of rules and measures used to regulate who or what can access specific resources within an organization. In the context of AI, inadequate access controls mean that systems may be exposed to unauthorized users, leading to potential breaches or tampering.

Read full story at The Decoder

More briefs