David Schwartz, Ripple’s CTO emeritus, warned that government rules on AI safety and copyright could grant authorities unprecedented control over political and social speech. He argued that two legal fights—one over whether training AI on copyrighted material counts as infringement, and another over AI safety regulation—could allow lawmakers to decide what the most powerful speech-generating tool can produce and discuss. Schwartz described this as “a totally unprecedented threat to freedom of speech.” His comments expand a recurring debate on X, drawing parallels to similar concerns raised by Erik Voorhees and contrasting with industry figures like Brian Armstrong, who opposes new AI-specific laws.
David Schwartz, Ripple’s CTO emeritus, has warned that government rules around AI safety and copyright could give authorities unprecedented control over political and social speech. His argument centers on who gets to decide what AI systems can create, know, and discuss.
Schwartz laid out the case in a reply thread on X that started almost by accident. On August 18, he quote-posted a Change.org UK campaign asking the British government to limit how much coverage any one person can receive in the press.
Schwartz’s response was a question: which poses more of a threat to free speech, artificial intelligence or natural stupidity. When a user named Athena asked what he meant, Schwartz explained that AI has become the most effective tool for producing speech, and governments are fighting over how much they can regulate it.
He spelled out two specific fights. The first is over whether training an AI model on copyrighted material counts as infringement, which would let Congress decide what speech looks like. The second is over AI safety regulation itself, following the same logic — if government sets safety rules, the most powerful tool for producing political and social speech becomes usable only in ways officials allow.
“Again, a totally unprecedented threat to freedom of speech,” he wrote.
This is not the first time Schwartz has pushed this line of thinking. In late July, Bitcoin advocate Erik Voorhees argued states should not decide what forms of intelligence count as safe, warning a rule against discussing bioweapons could expand into a broader government veto. Schwartz replied with agreement.
Telegram founder Pavel Durov, already facing terrorism-related charges tied to content on his messaging app, drew a fresh international arrest warrant from Russia’s security service in late July. The debate extends to proposals for government-backed AI testing after Google DeepMind CEO Demis Hassabis previously proposed a federally backed body to test models. Coinbase chief Brian Armstrong took the opposite view, contending existing laws covering fraud, torts, and consumer protection could already address harms.
