Mustafa Suleyman, CEO of Microsoft AI, believes that freely accessible online content can be used to train AI.
Mustafa Suleyman made the statement to CNBC at the Ideas Festival in Aspen, where he was asked whether AI companies were violating intellectual property by using data from the web to train their large language models (LLMs) without asking permission from or paying compensation to those who created them.
“As for content found on the open Internet, the social contract in force since the 1990s provides that it is subject to the principle of fair use (fair use),” explains Mustafa Suleyman. “Anyone can copy, recreate or reproduce it. It can be considered freeware. In any case, that’s our interpretation. »
At the same time, he acknowledges that things become more complicated if companies or news organizations declare that their content cannot be collected or copied. “We then find ourselves in a gray area on which the courts will have to decide. »
Let things be clear: Mustafa Suleyman’s statements are above all opportunistic. They recall those of Mark Zuckerberg, the founder of Facebook, who once said that the protection of privacy was a concept on the verge of extinction. In the same way, Mustafa Suleyman tries to make us forget that the economic model of many AI applications is based on data used without authorization.
And although many sites do not have detailed legal notices, they too are covered by copyright and their texts cannot therefore be copied without authorization.
It is also necessary to distinguish between commercial use and non-commercial use. Think, for example, of a teacher who finds Datanews.be articles so fascinating that he broadcasts them in his class. He does not risk any trial. But if you copy articles and then place them on your own website – after having rewritten them or not – the Roularta legal department may ask you, in a friendly but firm tone, to put an end to this practice. .
Mustafa Suleyman also seems to draw a distinction between those who can afford to sue and those who can’t. In recent months, OpenAI, in which Microsoft is a major investor, has struck deals with publishers including The New York Times and Axel Springer (publisher of Politico, Bild, and Business Insider, among others). The company is thus implicitly acknowledging that platforms whose content can be read for free should be compensated. But it apparently has no intention of paying for everyone’s content on the internet.
