I tested Gemini as my new assistant on Pixel - and it didn't go well
I've been using Android since its first iteration and I remember those early releases felt like they weren't quite ready for the masses. They were clunky, poorly designed, bereft of certain features, and prone to stalling or freezing.
Eventually, Android found its footing and has become one of the most flexible, user-friendly mobile OSes on the market. That doesn't mean Google hasn't been prone to making a misstep here and there.
Also: March Pixel feature drop delivers 10 exciting upgrades to Google's phone and watch
Such is the case with the migration from Assistant to Gemini.
If you've not been keeping up with Google's intention, its goal is to evolve the digital assistant that's been a part of Android for years to an AI-driven technology. What was originally called Bard was renamed and, in a very short time, was rolled out to some Pixel devices.
The idea was to make this change as seamless as possible. Instead of forcing users to dramatically change how they work with their devices, they could still use the tried and true, Hey, Google hot phrase to launch the AI-centric assistant, only they'd enjoy far more power. After all, isn't that the heart of AI?
Also: How to use Gemini: Everything you should know
Yeah, that's a trick question.
So, Google did its Googliest with Gemini, and some Pixel phones were given the gift of an AI assistant.
Naturally, because I write about this stuff, I was happy to give it a go. I assumed if Google was rolling it out, it was ready for prime time.
Funny that… Almost immediately, I discovered Gemini wasn't quite as capable as Assistant. I tried to create a calendar appointment, only to be informed that Gemini wasn't capable of connecting to the Google Calendar.
Send a text? Nope. Compose an email? Are you kidding?
At the time, Gemini could answer questions and take care of some very basic tasks, but it simply wasn't nearly as capable as Assistant. Granted, bit by bit, with each update, Gemini seems to be picking up more features but it's still lacking.
Also: I asked Gemini and GPT-4 to explain deep learning AI, and Gemini won hands down
For example, I used to be able to create Calendar appointments with Assistant. Now, when I attempt to do the same thing, I get a warning with an explanation that doesn't quite make sense. The warning tells me to tap on my avatar in the app, choose Extensions in the drop-down menu, and turn on the Workspace extension.
It took me a while to realize that message was referring to the Gemini app. I'd become so accustomed to configuring Assistant from within the Google app that I immediately defaulted to that app. So, I opened the Gemini app, tapped my profile icon, tapped Extensions, and enabled Google Workspace.
Why was that not enabled by default?
Ah, but there's a bit of a snag here. Let me backpedal for a bit.
One of the reasons why I've stopped using Google Drive for all of my work is that I don't necessarily trust that Google (or some third party) won't use my work to train its AI models. Given that I'm also a novelist, keeping my work safe from such things is rather important to me.
So, when I go to enable the Google Workspaces extension, I see that will "Summarize, find, and get quick answers from your own content to fuel your personal journey".
Also: I've published 60 novels - and these 5 free writing tools helped make it possible
That summary doesn't exactly instill confidence in me that Google isn't going to use my content for its own purposes. And because the Google Workspaces extension requires access to Gmail, Drive, and Docs, I'm most likely not going to enable the extension. Because of that, I'll miss out on features that were once part of Google Assistant.
How is that okay?
Essentially, Google is saying that, unless you give our AI access to your content, you won't get the most out of the new Android assistant. I guess I'm out of luck.
Anyone who takes their privacy seriously and doesn't want AI to have access to the totality of the things they've saved to Google Drive, written with Google Docs, or communicated with Gmail, will find themselves in the same boat.
Sadly, this limits Gemini (currently) to Google Flights, Google Hotels, Google Maps, and YouTube, which is pretty limiting. I'm guessing, if I leave the Workspace extension disabled, I'll find Gemini less than helpful and I'm just going to have to be OK with that.
Also: ChatGPT vs. Microsoft Copilot vs. Gemini: Which is the best AI chatbot?
The silver lining to this is that I tend to default to manually creating appointments in Google Calendar and have no desire to allow AI access to my content. In the end, if I want to stick with Android, I'll have to come to grips that it might be less efficient for me in the future. And the more embedded Gemini gets, I'm guessing the more access it will require.
Maybe I'm wrong. Honestly, I hope I'm wrong because I'd hate to have to switch to iOS at this point.