Anything posted online eventually turns into “freeware,” which can be duplicated and used to train AI models, according to the CEO of Microsoft’s new artificial intelligence (AI) business, in a recent interview with CNBC. He was asked if AI firms have truly taken intellectual property worldwide or not.

In response, Mustafa Suleyman stated, “Since the 90s, the social compact governing content currently available on the internet has been that it is fair use. It is freely replicable and repeatable by anyone. If you would want, that was freeware. It has been the mutual understanding.

When it comes to publishers and news organizations, Microsoft stands alone with AI companies:

He continued by saying that all content available on the internet can be used without restriction to train AI models, unless the content publisher or news organization specifically requests that Google refrain from scraping or crawling their content for purposes other than indexing, which is necessary to make content visible to other users. This has made Microsoft’s position on the usage of publicly posted content for AI model training quite evident.

By taking this action, the IT giant has aligned itself with AI businesses such as Perplexity, Google, and OpenAI, who maintain that training AI models on publicly available content is acceptable without requiring payment to the originator.

Disputes concerning the unauthorized use of content that is protected by copyright to train upcoming artificial intelligence models have consumed nearly every generative AI startup, including ChatGPT, Google Gemini, and Copilot.

Several organizations and media outlets, such as Forbes, the New York Times, and the Recording Industry Association of America, have launched legal actions against these corporations for reportedly utilizing their content to train their AI models without authorization, in addition to these controversies.