The US Secretary of Commerce emphasized the need to draft international legal frameworks that encourage innovation and enable model training without creating barriers for the next generation of developers.
Article Index:
- The US demand for the fair use principle
- The Department of Justice stance and major lawsuits
- Jensen Huang’s warnings against theoretical harms
- The joint G20 closing statement
- International competition and China’s participation in the consensus
- Frequently asked questions
The US demand for the fair use principle
The US administration urged G20 member countries on Wednesday to draft and develop unified international rules regulating how artificial intelligence companies use copyrighted materials to train their generative systems. This move coincides with mounting, complex lawsuits filed against major US tech companies in federal courts.
Leading companies such as Anthropic, OpenAI, Google, and Meta face widespread legal action from writers, publishers, news organizations, and artists, who accuse them of exploiting their works, books, and musical and artistic creations to train smart models that generate billions in revenue without obtaining prior approval or paying fair compensation. In response, tech companies defend themselves by asserting that training models by feeding them massive amounts of publicly available data constitutes a transformative and legally protected fair use.
US Secretary of Commerce Howard Lutnick asserted during a high-level ministerial meeting dedicated to technology in North Carolina that G20 nations must allow AI training on creators’ works while formulating a formula that guarantees the protection of artists. He stressed the need to adopt the concept of “fair use,” which permits the utilization of protected works under specific circumstances, without detailing specific mechanisms for achieving this complex balance.
The Department of Justice stance and major lawsuits
The Commerce Secretary’s remarks coincided with a highly significant legal step taken by the US Department of Justice on Wednesday, when it filed an official legal brief supporting OpenAI in its fierce legal dispute against The New York Times and several other major US newspapers, which accuse the company of exploiting their journalistic archives and protected articles to train the large language models powering the ChatGPT chatbot.
The Department of Justice argued in its brief that AI model training and development operations generally and primarily fall under the fair use umbrella established by intellectual property law, providing explicit and strong government backing for Silicon Valley companies against traditional content creators and established media institutions.
Jensen Huang’s warnings against theoretical harms
Jensen Huang, CEO of Nvidia, a leader in graphics processing and artificial intelligence, took the stage alongside Lutnick to issue a direct call to G20 officials urging them to avoid drafting restrictive legislation based on “theoretical and hypothetical harms,” inviting them instead to focus on addressing practical and tangible real-world problems associated with the technology.
Huang’s message reinforced the general direction of the US administration and the demands of top executives who participated in the event, including Tom Brown, co-founder of Anthropic; Alex Karp, CEO of Palantir; and Sam Altman, head of OpenAI, all of whom emphasized the danger of restricting technological innovation with overly strict bureaucratic laws.
The joint G20 closing statement
The discussions culminated in the issuance of a joint statement by G20 nations on Wednesday afternoon, in which countries affirmed their intention to avoid drafting new, dedicated regulatory laws entirely specific to artificial intelligence.
The statement emphasized that any future regulatory controls should be limited solely to addressing “new considerations and phenomena” generated by the technology that are not already covered by existing laws and legislation, to avoid a negative impact on the profitability of the tech sector and protect companies from any slowdown resulting from burdensome government requirements.
International competition and China’s participation in the consensus
The final document bore the signature of the People’s Republic of China, even though Beijing had already taken steps earlier to implement a broad and comprehensive package of laws and regulations specifically for artificial intelligence within its domestic markets.
This international consensus reflects intense geopolitical competition to shape the legal standards for the upcoming digital economy, amid persistent US efforts to establish flexible legal concepts as an international standard protecting its pioneering investments.
Frequently asked questions
Question: What is the fair use principle demanded by Washington for AI training?
Answer: It is a legal concept that allows the use of copyrighted materials for research or transformative purposes without prior authorization, which Washington believes applies to model training.
Question: What is the stance of the US Department of Justice on the New York Times and OpenAI case?
Answer: The Department of Justice submitted a brief supporting OpenAI, considering that language model training generally constitutes fair use of media materials.
Question: What did G20 countries agree on regarding AI regulations?
Answer: They agreed to avoid enacting new dedicated regulatory laws, relying instead on existing laws except for newly emerging technical cases and phenomena that have not been previously regulated.
