Generative AI
How AI models are trained and why it is a privacy risk.
🎯 Learning Objectives
- ▸ Understand how Generative AI works
- ▸ Analyze the privacy risks of training data
The Training Data Problem
Generative AI models are trained on billions of web pages, books, and images. If a blog post contains your name, email, or health information, that data might be ingested by an AI model.
Under the DPDPA, collecting this data without consent or a legitimate use is a violation. The Act requires a lawful basis for *any* processing, including AI training.
If you scrape the internet to train an AI model, you are processing personal data. You must ensure you have a legal basis for this processing under the DPDPA.
📖 Glossary
- Generative AI
- Artificial intelligence systems that can generate text, images, or other data in response to prompts.
- Training Data
- The dataset used to teach an AI model how to perform its task.
🔑 Key Takeaways
- AI models need vast amounts of data to learn.
- If the training data includes personal data, the model may "memorize" it.
- The DPDPA applies to the processing of personal data for AI training.
📚 Further Reading
- DPDPA Section 4 — Government of India (2023)
Shilpi Kulshrestha
Shilpi is a corporate lawyer with over a decade of experience in data protection, company law, and regulatory compliance. She has advised 50+ startups and enterprises on DPDPA readiness, privacy gover...