1) What data is sent to OpenAI?
No client details get passed to OpenAI (ChatGPT) aside from your specific question. Digits has a proprietary query language; we prompt-engineered OpenAI's Large Language Model to translate the user's question into a query we can execute without sharing any client details.
To be explicit, Digits transforms natural language via OpenAI to build requests into Digits' proprietary API. In doing so, Digits ensures that no client details are ever sent to another third party.Â
Client details never leave our system, and 100% of the computation is done in our backend.
2) What encryption/security is used when sending to OpenAI?
We use security practices such as encryption (TLS) in transit alongside prompt obfuscation to ensure that nothing can be tracked back to you, your clients, or your clients' data.
3) Who owns the data that is sent or received from OpenAI?
OpenAI does not own any data sent to it and cannot use this data for their models or training data. It is legally obligated to prevent this.Â
4) Is the data used for training models with OpenAI?
No.
5) Do we have legal backing that the data won't be used for training?
Yes, there is legal backing that prevents OpenAI from using any data sent via input to their systems. No client details are sent to any third party.
Â
