OpenAI and Google ask the government to let them train AI on content they don’t own
1 min read
Summary
AI companies OpenAI and Google have called on the US government to loosen copyright rules to allow AI systems to freely access and learn from copyrighted content.
They claim such a move is vital for US national security and to prevent China from taking the lead in AI, and point out that Chinese developers will have unlimited access to data, including copyrighted material.
OpenAI and Google both argue that exceptions to copyright rules, allowing the limited use of copyrighted, publicly available material for AI training, have been critical for the development of AI systems to date.
They both suggest that such exceptions should be extended, but with compensation for rights holders, while also calling for more government investment in energy infrastructure to support growth in AI.