OpenAI’s Post

Reading some of the FAQs, my initial thought is that it’s just lip service to security. For example, why do I need to fill out a form to have my data excluded? I feel that for the most renowned AI company, it should be as easy as yes/ no toggle in whichever app sitting on top of ChatGPt. Unfortunately, while I’m very excited about the potential of #GenAi, I’m very uncomfortable about sharing sensitive data into any of the models.

What’s the over under? Open Ai, Microsoft, Apple racing to that racehorse song Cripple Creek. Go. Call the race!

How can AI learn about the world; without learning about individuals? Although it seems scary, AI will eventually need to process and recognize patterns in human behavior to generate real world solutions. I am not the expert, just posing this scenario out of curiosity.

If you're concerned about privacy and ChatGPT, there are private AI options out there to check out. You can build your own (which is not easy). Venice.ai, Brave's Leo and DuckDuckGo's new private AI offering are all options that are off the shelf solutions.

It's important to consider that for these LLMs to behave and assist us intelligently and efficiently, they *must* train on human data. There is no other way to do it, and synthetic data training can only go so far. Quality ML-driven assistance = training on quality human-generated data. If you are bothered by your data being trained on, simply opt out of sharing anything private to you on the platforms by not entering it in the first place.

the models are black boxes. the black boxes have a risk of gaining agency themselves. for example, assuming goal-directed behaviour while creeping through networks. so, is it actually up to the company? should it be up to the/a company? asking from a reasonable place.

It's inspiring to see a commitment to user privacy! We'd love to learn more about the innovative measures you have in place.

Immediately after Elon Musk's comment on the Apple+OpenAI collaboration as "Cheap Spyware"

Thanks for sharing