Hackers could poison AI models by tampering with core infrastructure

T2Covered by 1 source · first reported by Cybersecurity Dive at 29 Oct, 14:27 UTC

AI systems rely on a complex web of hardware. Compromising any one component could let adversaries change model weights.

security

Sources

• T2Hackers could poison AI models by tampering with core infrastructureCybersecurity Dive29 Oct, 14:27