So Apple is saying: I will give Siri AI unrestricted access to user data and app because I trust my AI is good enough that it does not need a safety framework (permissions, visibility, accountability) to protect user data. But while Siri AI is safe (trust me bro), other AI provider are not so we will block them because the same unrestricted access we give Siri AI would cause issues. You see how delusional that sounds?
Reality, step by step, is:
Siri is trash
Please Google, give me an AI model that is good enough
Creating a safe environment for the model is hard, takes a lot of time, and the experience is bad for the user as security add too many blockers to the features
I can bypass all this complexity by just telling the AI not to do unsafe stuff
I cannot give third party access in this state, and I do not want to do it anyway as I like to keep total control like I do with App store. EU sucks
Apple seems to think that they have made Siri AI safe enough. We’ll see if that is true. Apple is sure taking a risk here.
If Siri AI is not safe, then the idea to make third party assistants safer would not work either (by the sound of it).
So Apple is saying: I will give Siri AI unrestricted access to user data and app because I trust my AI is good enough that it does not need a safety framework (permissions, visibility, accountability) to protect user data. But while Siri AI is safe (trust me bro), other AI provider are not so we will block them because the same unrestricted access we give Siri AI would cause issues. You see how delusional that sounds?
Reality, step by step, is:
Not really. Different systems have different capabilities.