Scott Farquhar thinks Australia should let AI train for free on creative content. He overlooks one key point | Artificial intelligence (AI)

According to Scott Farquhar, the founder of Atlassian, Australia must adopt the US -style copyright law to allow artificial intelligence to absorb all creative content or damage the investment that damages the sector in Australia.

Australian CEO technical council Farquuhar, Described ABC’s 7.30 program On Tuesday: “The entire use of mining or data search is probably illegal according to Australian laws and I think it hurts too much investment of these companies in Australia”.

The reason for this is that Australia does not have fair use exemptions coded to the copyright law like the US.

Farquuhar’s claim thinks that this is not a settled issue in the US and that it may have destructive effects on creative industries.

Companies that develop AI such as Atlassian, Google and Meta require a text and data mining exemption that is subjected to copyright law to ensure that AI can educate all human works in continuity.

_{Registration: AU Breaking News E -Post}

Farquuhar’s argument is that people are not theft of people’s work unless they are used to “directly copy an artist”, such as creating a song in his styles.

“I think people will say that, hey, if people will sit with a digital friend, AI song creator and work in cooperation with an AI to create something new to the world, which is probably a fair use.”

Farquuhar, the benefits of large language models, AI’nın other people’s data about the work of free of charge, he said.

His argument depends on whether AI’s copyright, called a transformative – that creates something new, creating something new.

He said he wouldn’t have a problem with his use of what he created with someone and used it as long as he is “transformative.

“If someone had used my intellectual property to compete with me, then I think it was a problem with me. If he used all the intellectual property of all software in the world to help people write software in the future, I think this was a fair use.”

The US law has not been solved to be used in a fair use of AI education. USA copyright office recorded In the front raid in May In productive AI education, there are dozens of cases to AI companies, AI companies, as a excuse to educate large language models on copyright -protected works.

In the USA, there are factors that should consider whether something is a fair use:

• Whether the use is commercial

• NATURE OF THE RIGHT RIGHT

• The amount of use of the copyright protected work

• The effect of use on the market or value of the work -protected study.

In the US case -law, the transformative nature of the things taken, the first important factor in fair use, Australian law firm Gilbert + TOBIN Was noted in May However, the effect of the work on the market is the key. According to the US Copyright Office report, the US Supreme Court described it twice as “the only one of the most important fair use of fair use”.

“Copying in artificial intelligence training threatens significant potential damage to the market or value of the copying, copy -protected studies,” he said.

“In cases where a model can produce significantly similar outputs that replace the studies in training data, it may lead to the loss of sales.

The outputs of a model can dilute the market for similar works, including the production of similar materials, including the production of similar materials, even if it is not largely similar to any copyright. “

The Copyright Office stopped recommending the legislative intervention in the USA, and voluntary licensing that it has already continued with some AI companies and allowing licensing, AI innovation will allow “intellectual property rights without weakening”.

If Farquuhar has appropriate guarantees that all use will be a transformative and will not affect the markets they draw, the argument of how he uses artificial intelligence that produces copyrights.

In many industries, AI may have devastating effects. In the news, for example, the AI summary in Google seary means that people are less frequently clicked for information and that the instructions made from AI chatbots are even worse once AI scans a page.

To claim that fair use for artificial intelligence based on the US law is something that Australia should desire to look at it is almost no established law and that it is forced to fight in the courts. Hurry to give technology companies what they want for innovation for a new industry may be at the expense of many other industries.