Orca: AI services, models falling short on security

Orca Security’s “2024 State of AI Security Report” found organizations failing to apply basic security measures to AI tools and models. Many firms struggle to disable risky default settings that could be exploited by threat actors. The findings reveal significant security shortcomings as AI adoption continues at pace. Software packages used by 56% of firms deploying their own AI models contained at least one high-risk vulnerability. Orca urged firms to check default settings and implement strict security policies to alleviate the risks.