Former DeepMind executive Verity Harding warned that the current geopolitical tone around AI is increasing the risk of dangerous outcomes. She pointed to nationalistic policy signals as evidence that governments may be treating AI development as an arms race.
The concern is that competitive pressure can weaken safety coordination, especially if major countries prioritize speed and strategic advantage over governance. That dynamic has become more visible as AI policy moves into national security debates.
The interview adds a governance perspective to the technical race, emphasizing that deployment incentives may matter as much as model capability.