Al Jazeera questions who should set global AI speed limits

This column captures a growing unease about who actually gets to steer the global AI timeline. When Anthropic’s Dario Amodei and peers call for a coordinated slowdown of frontier models, they present it as a safety intervention. The piece argues that such coordination would also consolidate power among a handful of US-based firms, potentially carve out antitrust exceptions and explicitly enshrine export-style restrictions on China’s access to the highest-end chips and compute. That is not just about safety; it is about who writes the rules of the emerging AI economy and who gets locked out.

For the race to AGI, the essay highlights that “speed limits” set by incumbent labs and Washington could entrench US platforms while leaving China and the rest of the world reacting on unfavorable terms. It also points out that energy-rich regions such as the Gulf could use control over power, land and permitting for massive data centers as bargaining chips, demanding a real say in governance rather than accepting frameworks negotiated in Washington, San Francisco and Beijing. If those regions insist on representation and attach conditions to hosting AI infrastructure, the de facto timeline for scaling compute and models could be shaped as much by geopolitics and energy policy as by technical feasibility.