The AI accord leaves weapons out

The White House Accord on Super Intelligence, signed on September 29, sets out safety checks for AI models but leaves autonomous weapons and cyber agents largely outside its scope, The Express Tribune argued. The proposed measures include internal safeguards, an oversight team and independent audits during model development and use, aimed at risks involving cybersecurity, biological threats and chemicals.

The article says the accord does not impose a legal duty on signatories to act, nor does it cover AI-enabled drones or agents capable of launching complex cyberattacks. It also questions whether an agreement between the United States and China is achievable while their AI capabilities remain unequal, noting that earlier arms-control deals were reached between strategic peers or after one side accepted limits from a weaker position.

The commentary sees commercial competition and national security concerns as obstacles to restraint, while arguing that cooperation could still be possible if both governments prioritise reducing existential risks. It notes that China’s chief spy official has also warned of AI’s potential threat to political stability.