Why do AI companies do such malpractice?
Consumer side: All major AIs train on user data by default (you must opt out).
Business / API / Enterprise side: They do not train on customer data by default. You need to opt in if you want your data to be used to train models.
😡😡😡😡😡😡😡😡
This includes Sarvam AI too.
Knowledge bases are collections of your own uploaded documents (PDFs, DOCX, TXT, etc.) used for RAG in agents (Voice Agents, Cowork, etc.). These count as “uploads” / “Your content.”
Sarvam says: “We use Your content (including inputs, uploads, prompts, or generated outputs) to train, fine-tune, and/or improve our AI models unless you explicitly opt-out through your account settings or by writing to them.”
Section on “AI Model Training & Your Data” states the default policy is Opt-In.
