Tech Giants' Data Hunger Drives Questionable Tactics
-
Tech companies like OpenAI, Google, and Meta are desperately seeking more data to train their AI models, even if it means bending rules or laws.
-
OpenAI created a tool to transcribe YouTube videos without permission to get more text data, while Google may have done the same.
-
Meta executives discussed buying a publishing house and using copyrighted online content for their models, arguing it qualifies as fair use.
-
The tech giants have nearly exhausted reputable English language text online but need more to advance their AI.
-
Facing a data shortage, companies like OpenAI are exploring using AI-generated "synthetic" text to train more powerful AI systems.