This latest Pixel drop brought one of my biggest fears to Android: notification summaries. It’s something Apple botched in iOS 18 last year, and something I didn’t feel like Android needed given its top-tier notification management. After a week with the feature, though, I’m of two minds. Google is definitely doing this the “right” way, but does that mean I’m likely to leave this tool enabled for weeks to come? Unfortunately, not quite.
Let’s start with the good. In many ways, Google has clearly learned from a year’s worth of bad press scattered around Apple’s implementation of this feature. That’s an interesting case study; Apple tends to be the company that is “late” to the party, with the excuse that any time it is late to something, it gets it right on the first try. That was absolutely untrue with notification summaries, given their general bugginess, inaccuracies that bounced between hilarious and horrifying, and months of missing support for news apps after too many embarrassing slip ups.
Google, instead, is going for a far more manageable route. Notification summaries on Android are meant to only work with messaging apps in their current iteration. This support is fairly widespread — every messaging app installed on my Pixel 10 supports it — so if you’re interested in this tool, you know you’re going to get to use it no matter where you’re chatting from. It’s also focused on just summarizing group chats and long messages, so in theory, it’s only working when a summary would actually benefit you.
Of the summaries I’ve seen over the past week, most of them have been pretty accurate, correctly taking the language within any given notification and properly summing it up in a quickly digestible card. These notifications are clearly delineated as an AI-powered summary — full italics support a la Apple, along with AI’s tell-tale sparkle emoji. Because you’re only summarizing incoming messages, I think this is pretty vital to the feature’s success. If it’s going to be scattered among normal, non-AI alerts, it has to be obvious.
Crucially, it also keeps things in first-person — at least, most of the time. A seasonally-appropriate text from my fiancée alerting me that the pie she’s baking will be done in the next hour, should I want her to run to the store, was correctly summarized and kept her use of “I,” meaning the alert still felt fairly personable. That’s a far cry from the third-person all-observant tone Apple’s notification summaries used in my experience, which felt cold and separate from the actual friend or loved one I was speaking with. It’s a good way to keep things human, which, above all else, is one of my biggest pet peeves with my AI experience to date.
Unfortunately, while Google has fixed plenty of the notification summary issues I found on iOS, it has failed to find an actual reason for me to keep these enabled. That pie-baking example might look good at first glance, but — and I manually counted this — it only saved 23 characters total off the original text. Does that save me any time? Does that deliver a more straightforward, streamlined thought? Not really, if I’m being perfectly honest.
Other examples from the past week are similar. Sure, a text about an upcoming delivery from Walmart is shorter in its AI-summarized format, but to what end? The full alert would’ve still been seen in my notification tray had I kept this feature off. Likewise, recent The Sideload podcast guest Ara Wagoner asked me over Instagram DMs if she needed to locally record her audio or if she could utilize her Chromebook, and while the AI did get that point across, it reads a bit jumbled — and, curiously, drops that first-person POV I just praised.
It can also fail to recognize mistakes in text messages, something that we humans are bound to do whenever we’re typing too quickly for our brains to keep up. One of my friends, in the midst of asking about upcoming Friendsgiving plans, accidentally sent a text to myself that was meant for one of his co-workers. I would’ve recognized this mistake immediately — and did, for the record — but Google’s AI summary of this conversation did not. Instead, it’s included in a round-up of these upcoming messages, and one that makes it more difficult to see that slip-up in the first place.
Unsurprisingly, Google’s summaries are probably best at trying to condense productivity or work-focused messages, particularly in a group chat. On paper, that makes it prime to shrink down some of the incoming Slack messages I might get throughout the day. And yet, I’ve really only seen it work once in this method, summarizing a couple of to-do messages shared between fellow 9to5 writers Andrew Romero and Ben Schoon. It took nearly the entire week for this example to pop up in my notifications, and unfortunately, it’s the only time where I felt actually assisted in its wake.
Google’s implementation here is undoubtedly better than Apple’s, even with the latter having a year’s head-start. And yet, I don’t feel like my Pixel 10 has proved to me that this tool really serves a purpose. Maybe it’s my always-connected state that simply renders notification summaries pointless, but I just can’t see the vision here. Android’s notification system is so good at presenting me what I need to know, trying to lay AI on top of it feels superficial at best, and superfluous at worst. While I’m sure this is just the beginning of what Google has planned for this feature, I think most people would be best served by keeping this disabled on their devices.
FTC: We use income earning auto affiliate links. More.
Comments