David Schwartz, Ripple’s CTO emeritus, has voiced significant concerns that government regulations targeting AI safety and copyright could grant authorities unprecedented control over political and social speech. Schwartz articulated these worries in a series of posts on X, formerly Twitter, suggesting that the frameworks being developed to govern artificial intelligence might inadvertently empower officials over public discourse.
Schwartz’s argument focuses on two key regulatory areas: the use of copyrighted material for AI training and the definition of AI safety. He explained that if training AI models on copyrighted content is deemed copyright infringement, Congress, which controls copyright exceptions, could dictate the permissible uses of AI speech-generating tools. This scenario, according to Schwartz, would effectively allow lawmakers to shape the nature and scope of AI-generated speech.
Furthermore, Schwartz warned that government-defined AI safety rules could pose a substantial threat to free expression. He posited that if the government establishes the criteria for what constitutes ‘safe’ AI, then the most powerful tools for producing political and social speech could become usable only in ways that officials permit. This, he argued, represents a “totally unprecedented threat to freedom of speech.”
These concerns were initially prompted by Schwartz’s reaction to a Change.org UK campaign advocating for limits on press coverage for individuals. This campaign gained traction following extensive media attention on sociology professor Jason Arday, who was the subject of 249 articles in the 22 days preceding his death on August 14. Schwartz questioned whether AI or “natural stupidity” posed a greater threat to free speech, sparking a discussion about AI’s role in generating speech.
Schwartz elaborated that AI has become an exceptionally effective tool for producing speech, and governments are currently navigating multiple regulatory challenges concerning its use. He echoed sentiments previously expressed by Erik Voorhees, founder of ShapeShift, who argued in late July that governments should not be the arbiters of what forms of intelligence are deemed safe, cautioning that such power could expand into a broader veto over speech and encryption.
The broader debate over AI governance and its impact on speech is not isolated. Proposals for federally backed bodies to test and certify advanced AI models, such as one suggested by Google DeepMind CEO Demis Hassabis and supported by OpenAI CEO Sam Altman and Microsoft’s Satya Nadella, have also surfaced. However, figures like Coinbase chief Brian Armstrong have suggested that existing laws related to fraud, torts, and consumer protection might be sufficient to address potential AI-related harms.
Schwartz’s perspective highlights a critical uncertainty: who ultimately decides what AI systems can create, know, and discuss. The potential for AI safety regulations and copyright law interpretations to grant governments significant oversight over digital communication remains a key concern, with the implications for political and social discourse yet to be fully determined.
Broader Context
Source materials place the factual news in this context: 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.
Leave a Reply