Microsoft AI CEO Mustafa Suleyman made somewhat controversial statements when he said in an interview with CNBC that everything published on the internet is “freeware” — that is, anyone can copy and use it as they wish. The statement came after the interviewer’s question, who wanted to know whether AI companies take intellectual properties from the web to train artificial intelligence models.
AI can be trained with everything on the internet
When calling online content “freeware,” Suleyman says there is no problem using network publications in generative AI training, such as OpenAI’s ChatGPT and Microsoft’s Copilot.
“I think that, with regard to content that is already on the open web, the social contract for this content since the 90s has been fair use”, points out the CEO of MS. “Anyone can copy it, recreate with it, reproduce with it. This has been ‘freeware’ if you prefer, that has been the understanding”, he concludes.
The problem with this “understanding” is that there are countless cases of intellectual property online, from videos on YouTube to texts on sites. Furthermore, photos and other media also fall under different rules, just like content on social networks. Therefore, it is likely that Suleyman is wrong when he claims that everything is “freeware” on the internet.
Microsoft and OpenAI are being sued
It is worth remembering that in December 2023, the American newspaper The New York Times filed a lawsuit against Microsoft and OpenAI, accusing the companies of copyright infringement in training artificial intelligence models.
The famous news portal claims it lost “billions of dollars” because chatbots used its content and information without permission.
Source: CanalTech