Artificial intelligence has been a source of division for years. Since accessible generative AI arrived in 2022, institutions, researchers, students, activists and policymakers have argued over its ethics. Now the disagreement has reached the executives who build these systems.

Dario Amodei, chief executive of Anthropic, published an essay urging AI companies to slow work on their most advanced models, arguing that existing safeguards cannot keep pace. His proposal would restrict China's access to the most powerful AI chips and advanced semiconductor manufacturing equipment, and curb smuggling and remote access to computing abroad. China's foreign ministry rejected the proposal as fearmongering aimed at holding back its technological progress, invoking Cold War parallels. Amodei himself acknowledged that China poses the hardest problem for his plan.

The industry is not united. Jensen Huang, chief executive of Nvidia, which sells the chips these companies rely on, opposes a slowdown and argues that market forces, not new legislation, can push firms toward safe innovation. Mark Zuckerberg, whose company releases its models for free, warned the proposal could expose companies to greater legal risks. David Sacks, the White House AI lead, has criticized the idea of letting AI companies work outside antitrust law, questioning how one company can fairly inspect another.

Supporters have compared coordination to banking regulation after the 2008 crisis, but Basel III followed an actual crisis and was imposed by outside regulators, not by competing firms coordinating among themselves. The deeper question is not only how fast AI should move, but who gets to decide.

Source: Al Jazeera