100% agree with your points on AOSP. I was going to add a lot of other stuff about that to my rant but I thought it was long enough already. The biggest worry, IMO, is the proprietary part of Android (Google Services) becoming a hard requirement by more and more apps in the name of “security” and “certification”.
Anyway, time will tell and people will find a way forwards
I quite agree with most of you, AI is not just an innocent software that should be implemented by default and without notice (firefox’s way). It has several impacts on ourselves and the society in general.
I just thought that I had immediately discovered the switch (desactivated by default?) after the upgrade to 7.3 but now discover another switch that was activated!! Both are in “notifications”
It’s neither stupid nor irrational in a privacy-focused operating system. Few of us have the background to analyze new software and figure out how it works and what servers it’s sending data to and how it may be violating our privacy. Therefore, I don’t want AI or anything similar to it. If I did not care, I would still be using Scroogle and normal Android phones.
Does that actually completely turn all the “A.I.” features off?
I see the top two items in this list as being AI-like because these features observe user behaviour and make internal changes in response. However, I don’t see how the bottom feature would qualify as AI because it is a simple setting that is on or off - it does not make internal changes based on any learning of user behavioural patterns.
AI is a “trigger-word” for me. I appreciate this conversation discussing the nuances of what the term means and implies, and how it applies to Iode.
As I understand these “smart notification features” (note I continue to object to them being termed “AI”) are local only there are no servers involved.
putting aside the serious implications of heavily integrated non-consensual data recording, even if it’s intended to stay on your device. (I don’t like the iPhone taking frequent IR snapshots of my face even if it stays on my device, for several reasons that should be apparent, even more so if it’s not disclosed to me)
the main over arching issue is not having control over your software and hardware and personal data. When there are decisions made for you on your device that are not explicitly defined in the source code, then you don’t have control over the software.
You can see a decision tree in source code. And with properly annotated source code you clearly can see under what circumstances decisions will be made
However people very often cannot see under what circumstances decisions will be made with AI. The AI may be open source but many contemporary types of “AI” being heavily integrated today do not have a completely defined set of circumstances that decisions will implicitly be made under with high certainty.
The developer may not even know the complete set of circumstances much of the time. Which is why id argue it’s a bad practice in far too many cases that it’s used for.
People choosing open source software so they know what their software will do should not be gaslit when they say “AI” in the source code is preventing them from knowing, especially when in many cases the developer even doesn’t know. (I’m not accusing anyone here of gaslighting about this)
it’s worth saying too that not knowing all of the circumstances under which an AI decides how a npc moves in a video game is very different from not knowing all of the circumstances under which decisions are made by your operating system for any of the many the important functions it performs especially with sensitive data.
You might respond to a lot of this to just say “most users don’t even read the source code”. Yea but many of them take the advice of the people who do read the source code seriously to figure out what their software will do. And when people who actually do read the source code can’t successfully figure out under what circumstances decisions are being made by the operating system then there’s clearly a problem for everyone who wants to know what their software will do.
yes, we can, and after hearing people’s concerns, the team has decided to implement this feature as disabled by default in the next beta:
The beta for iodéOS 7.4 will be out soon, with (among many improvements) the Notification Organizer disabled by default. We will include documentation on how to enable it for those of you who do want the feature in the future. For the moment, people wishing to disable it in 7.3 can simply go to
Settings > Notifications > Notification organizer
and uncheck “Use Notification Organizer”