AI Safety: Balancing Control and Equalization? // Kyle Harrison // MLOps Podcast 180 clip

MLOps podcast #180 with Kyle Harrison, General Partner at Contrary, The Centralization of Power in AI.

The main threat in AI is the concentration of power. Closing off AI models limits visibility and control to maintain power and distribution. This, in turn, creates a vulnerability where many entities become dependent on a single point of failure. The argument against this centralization is that it doesn't prevent bad actors from accessing technology, as they eventually do. The better approach is to …

ai models clip control distribution equalization general general partner mlops partner podcast power safety threat visibility vulnerability