Mustafa Suleyman, the co-founder of DeepMind and CEO of Microsoft AI, recently spoke with CNBC about a contentious issue. Suleyman stated that any content on the web can be used to train AI without seeking permission from the owners. He mentioned that content that is freely available on the open web has been considered fair use since the early days of the internet in the 90s. This includes copying for use, creating new works, and utilizing it as freeware.
However, there are exceptions where content specified on websites or by publishers explicitly states that data extraction for any use is prohibited, except for indexing purposes. Suleyman views this as a gray area that may require legal intervention to find a resolution.
The debate about using internet data to train AI remains a topic of contention, with differing viewpoints among parties. Examples like OpenAI and Microsoft have faced legal challenges when agreements with news agencies were violated.
Source: Windows Central
TLDR: Mustafa Suleyman, the co-founder of DeepMind and CEO of Microsoft AI, discussed the use of web content for training AI models without permission, citing it as freeware according to the “social contract” of the internet.
Leave a Comment