OPENAI and Google ask the government to let them form AI on the content they do not have
OPENAI and Google push the US government to allow their IA models to train on material protected by copyright. The two companies have described their positions of proposals published this week, with Open disputes that the application of the protections of fairness to AI “is a question of national security”.
The proposals come in response to a request from the White House, who asked governmentsIndustry groups, private sector organizations and others for contribution to the “AI action plan” by President Donald Trump. The initiative is supposed to “improve the position of America as a power of AI”, while preventing the “heavy requirements” of having an impact on innovation.
In his commentary, the open claims according to which authorizing the companies of AI to access content protected by copyright would help the United States Call the rise of Deepseek.
“There is no doubt that the RPC [People’s Republic of China] IA developers will benefit from unhindered access to data – including copyright -protected data – which will improve their models, ”writes Openai. “If RPC developers have unhindered access to data and American companies are without equitable access, the AI race is over.”
Google, not surprisingly, agrees. THE Business response Likewise, says that copyright, confidentiality and patent policies “can hinder appropriate access to the data necessary for the training of head models”. He adds that the fair trade policies, as well as the exceptions of text and data exploration, were “essential” in the training of AI on data accessible to the public.
“These exceptions make it possible to use material protected by copyright for training in AI without having a significant impact on rights and avoiding negotiations often very unpredictable, unbalanced and long with data holders during the development of the model or scientific experimentation,” explains Google.
Anthropic, the company of IA behind the chatbot ai claude, also submitted a proposal – But that does not mention anything of copyright. Instead, he asked the US government to develop a system to assess the national security risks of an AI model and strengthen export controls on AI fleas. Like Google and Openai, Anthropic also suggests that the United States is strengthening its energy infrastructure to support AI growth.
