OpenAI urges Trump to permit AI training on copyrighted content
What's the story
OpenAI, the company behind the famous ChatGPT chatbot, has made a set of recommendations to the current US administration led by President Donald Trump. The main goal is to push for deregulation and policies that would allow AI companies to use copyrighted content freely in their model training. The move comes as an effort to strengthen competition with China in the race of AI development.
Regulatory concerns
OpenAI argues against restrictive state laws
OpenAI has voiced concerns that restrictive state laws prevent AI companies from innovating freely. The organization argues that these regulations are "easier to enforce" on domestic firms and impose compliance obligations that "weaken the quality and level of training data available to American entrepreneurs." To combat this, OpenAI is proposing a "copyright strategy" to preserve the ability of "American AI models" to learn from copyrighted content.
Fair use advocacy
Importance of fair use in AI training
OpenAI argues AI models should be trained on copyrighted data as they're "trained not to replicate works for consumption by the public" and thus fall under the fair use doctrine. The organization claims restrictive copyright laws in the EU have stifled AI innovation and investment. Without fair use access to copyrighted data, OpenAI warns, "the race for AI is effectively over" and "America loses."
Data sharing proposal
Tiered system for AI data sharing
When it comes to AI data sharing, OpenAI has suggested a tiered system. This would enable sharing of AI tech with countries following "democratic AI principles," while denying access to China and restricting countries that could leak data to China. The company also calls for government investment in the application of AI tech and building AI infrastructure as part of its wider plan to promote innovation and competition in the field.