Advogado, você pode estar VAZANDO os dados do seu cliente com IA (e nem percebe)

Do you use ChatGPT, Gemini, or Claude in your office? Then you need to watch this. By sharing a process, contract, or document with AI, you enter the data processing chain as a controller and become subject to the sanctions of the LGPD (Brazilian General Data Protection Law) and the ANPD (Brazilian National Data Protection Authority). In this video, I show the real risks (including international data sharing and leakage of trade secrets) and, most importantly, the practices I use in my office to make this use much safer: disabling model training, NotebookLM, offline local AI, and intelligent anonymization. I'm Philipe Cardoso, a lawyer specializing in LGPD and Law 4.0. If the content helped you, subscribe to the channel so you don't miss the next updates. ⏱️ Chapters: 00:00 Introduction 00:27 How I use AI in my daily office work 01:20 Are you handling client data well? 01:57 Why the lawyer is a data controller 02:17 Sanctions of the LGPD and the ANPD 02:43 International data sharing 03:37 How to choose suitable companies (LGPD/GDPR) 04:34 How the ANPD determines the penalty dosage 06:02 AIs learn from your conversations (the risk of "red sky") 06:59 Trade secrets, trademarks and patents 07:22 Disable model training 08:11 NotebookLM: a safer environment 08:39 Running AI locally (offline) 09:27 Anonymization with black bars 10:01 Automatic anonymization with local AI 10:49 Conclusion