I can’t solve the ethical bit, but I am working on a project that can solve the first two – I.e., using large language models through API key access in a secure and private way. It’s challenging, but we’ve made some really good progress using a number of layered approaches. I plan on posting the project here once we get a better confirmation and testing on the privacy and security aspects.
Privacy and Security may also depend on the server where the AI model is runnig. Having a 3rd party tool using the API is a nice thing for sure.
But OpenAI will still use the prompt data and sell that (if someone inputs sensitive information. So privacy gone.
For security similar.
I can’t solve the ethical bit, but I am working on a project that can solve the first two – I.e., using large language models through API key access in a secure and private way. It’s challenging, but we’ve made some really good progress using a number of layered approaches. I plan on posting the project here once we get a better confirmation and testing on the privacy and security aspects.
Privacy and Security may also depend on the server where the AI model is runnig. Having a 3rd party tool using the API is a nice thing for sure. But OpenAI will still use the prompt data and sell that (if someone inputs sensitive information. So privacy gone. For security similar.
Don’t bother. Lemmy is too far up it’s own ass. You may be able to have an actual discussion on Hacker News.