gtag('config', 'G-0PFHD683JR');
Bitcoin

The United States Copyright Office has ideas about artificial intelligence. You may not like big technology.

Large technology companies are training artificial intelligence models often on the work of other people, such as scientists, journalists, directors or artists.

These creators have long objected to this practice. Now, it seems that the US Publishing Office has joined their side.

The office issued last Friday in a series of reports that explore the laws of publishing and artificial intelligence. The report deals with whether the copyright content that AI uses to train the artificial intelligence models are qualified under the doctrine of just use.

Perhaps artificial intelligence companies will not like what they read.

Artificial intelligence companies desperate data. Most of them believe that the more information that the model can target, the better. But with this inconsistent consumption, they risk running the laws of copyright.

Companies like Open AI have faced a large number of lawsuits from creators who say that training artificial intelligence models for their work with copyright without permission to violate their rights. Ai Execs argues that he did not violate copyright laws because training is under fair use.

According to the new report of the copyright in the United States, it is not so simple.

“Although it is not possible to judge in advance in any specific case, the precedent supports the following general notes,” the office said. “The various uses of the copyright -protected work are likely to be transformed. However, the extent that depends on fair, will depend on what has been used, from any source, for any purpose, and what are the controls on the outputs – all of which can affect the market.”

The office distinguished between the artificial intelligence models of research and artificial intelligence models.

The office said: “When a model is published for purposes such as analysis or research – types of uses that are decisive of international competitiveness – it is unlikely that the outputs will replace the expressive works used in training,” the office said. “But the commercial use of vast lines of copyrights protected to produce expressive content that competes with them in the current markets, especially when this is achieved through illegal access, exceeding the limits of fair use.”

In the report, the office compared the outputs of artificial intelligence that mainly copy training materials to the outputs with additional elements and new value.

The office said: “On the two ends of the spectrum, the training of the model is more transformed when the purpose is to spread it for research, or in a closed system that restricts it on a task that is not built.” “For example, a language model training on a large collection of data, including social media, articles and books, to publish in the systems used in the content of content, does not have the same educational purpose as these papers and books.”

Training the artificial intelligence model to create outputs “is very similar to the copyright of the data group” less likely to celebrate.

The office said: “Contrary to cases where computer programs were copied to reach their functional elements, it is necessary to create new operational works, using photos or audio recordings to train a model that generates similar expressive outputs that does not remove merely a technical barrier for productive competition.” “In such cases, unless the original work itself is targeted for suspension or satirical simulation, it is difficult to see use as a transformation.”

In another section, the office said it rejected two “involved arguments” about “the transformative nature of artificial intelligence training.”

The office said: “As mentioned above, some argue that the use of copyright works to train artificial intelligence models is a transformation of its nature because it is not for expressive purposes. We see this argument wrong.”

“We do not agree that the training of artificial intelligence is the transformation of its nature because it is like human learning,” he added.

A day after the office issued the report, President Donald Trump launched its manager, Shera Permotter, a Business Insider spokesman.

“On Saturday afternoon, the White House sent an email to Shera Permottter,” your position is as a record of copyright and director of the copyright office in the United States. “

While Trump, with the help of Elon Musk, who has his artificial intelligence model, Grok, sought to reduce the federal workforce and close some agencies, some have seen the time of Perlmturt’s dismissal as a suspect. New York MP Joe Morel, a democratic, addressed the launch of Berlipler in an online statement.

“The end of Donald Trump to record copyright, Shira Berlipreer, is a rude, unprecedented power without any legal basis. It is certainly not by chance that he acted less than a day after her rejection, the statement said.

Big Tech and AI have gathered around Trump since his election, led by Musk, which has become the face of the Douge Doge office and management efforts to reduce federal spending. Other technical billionaires, such as the CEO of Meta Mark Zuckerberg and the CEO of Openai, Sam -German, has also made Trump in recent months.

A White House actor did not respond to a comment from Business Insider.

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button