Cheap model APIs raise code privacy worries

Using subsidized directly can expose product code or product ideas to the companies that run the models. The main concern is that a feature being built today could appear in a future model release.

The worry is stronger for from Chinese companies. The practical question is how to keep code and private product details safe while still using lower-cost external models.

Key points

  • Low-cost may create code and product privacy risks.
  • The concern is that s could learn from sensitive inputs.
  • The privacy worry is stronger for Chinese model companies.
  • should weigh cheaper against data protection needs.
Read original