Learn
BigLLMs
Pinpointing some security issues with top LLM providers
When using services provided by large language model (LLM) providers, it is essential to be aware of the inherent security concerns. Typically, your sensitive data will leave your computer, be transmitted over the internet, received and stored, and then processed and analyzed.
Sensitive data may be shared with vendors and service providers, disclosed during business transfers, revealed to law enforcement or other authorities, and may be subject to international data transfer or jurisdiction laws.
ChatGPT OpenAI
Default Data Retention
- 30 days
Notable Breaches Incidents
Claude Anthropic
Default Data Retention
- Permanently in chat mode
- 30 days in API mode
Bing Microsoft
Default Data Retention
- 6-18 months
Additional Concerns
- Use for 3rd party marketing and targeted ads
- Data collection by advertising and tracking partners
- Data combination from third-party sources
Bard Google
Default Data Retention
- 9-18 months
Additional Concerns
- Use for 3rd party marketing and targeted ads
- Data collection by advertising and tracking partners
- Data combination from third-party sources