Anthropic has stated it has never advocated for banning open-weight AI models, instead focusing its concerns on misuse by authoritarian states, including cyberattacks and biological threats. Crypto figures like Erik Voorhees and Ripple’s David Schwartz have warned that allowing governments to define "safe" AI knowledge could create a dangerous precedent, expanding from bioweapons to unapproved encryption and beyond.
Crypto figures are warning that government control over what AI systems can know or discuss could create a dangerous precedent. This follows a post by Erik Voorhees on July 28, arguing states should not decide what forms of intelligence are “safe.”
Voorhees, a longtime Bitcoin advocate and founder of ShapeShift, posted after Anthropic CEO Dario Amodei published a statement. Amodei insisted his company has “never advocated a ban on open-weight models.”
The executive argued such systems provide value when they lack dangerous capabilities. Anthropic’s main concerns involve authoritarian governments developing AI for military or surveillance uses, along with misuse in cyberattacks or biological threats.
Anthropic supports restrictions on access to advanced chips and safety testing for highly capable AI systems. The Trump administration is reportedly finalizing a framework for companies to voluntarily submit advanced models for government testing.
Voorhees argued this could expand beyond its original purpose. He walked through a hypothetical progression from rules against discussing bioweapons, to banning unapproved encryption.
“Civilization must not permit the state to determine what manner of intelligence is ‘safe,’” he wrote. Ripple CTO Emeritus David Schwartz replied with “Yes, this. Exactly this.”
The debate also emerged when Google DeepMind CEO Demis Hassabis proposed a federal body to test frontier models. Coinbase CEO Brian Armstrong rejected the idea, stating existing fraud and consumer protection laws already cover potential harm.
