Google didn’t copy Liquid Glass. It did something even worse

Good news, everyone: Google hasn’t copied Apple’s Liquid Glass interface design after all, despite hinting at this in promotional materials. All the Android users who pushed back so vehemently will doubtless be pleased, although it’s debatable whether the thing it actually announced is much better. Instead of copying Apple’s surface aesthetic, Google has copied its unhealthy obsession with system-wide AI.

This year, the company has revealed, Android will transition “from an operating system into an intelligence system.” Its Gemini AI tech will be integrated at all levels: automating tasks across apps, offering autofill suggestions, converting natural speech to more polished text, and acting as a gatekeeper for web browsing. Whenever you turn on an Android device with the new OS, you’ll be confronted by a pushy AI assistant offering to do everything for you.

Google calls this setup Gemini Intelligence, which feels a bit on the nose. I’m not sure why it didn’t go the whole hog and call it Gapple Gintelligence. But a lack of originality is par for the course from Apple’s rivals. Google is just being unoriginal in an unexpected way.

To be clear, I wouldn’t for a second accuse Google of copying Apple’s AI technology. Apple came late to the AI party and still hasn’t caught up, whether because of wise patience or just a frustrating lack of R&D success. Right now, Gemini is objectively better than Apple Intelligence. What Google is copying–and where it’s making a strategic error, in my view–is Apple’s structural and presentational approach to AI. Which is to say, shoving it down users’ throats at every opportunity.

Some users, I suspect, will find this approach alienating: Most people simply do not love AI as universally and uncritically as C-suite tech executives. As things stand, the (admittedly significant) proportion of Android users who are fascinated by AI can seek out the relevant apps and services and play with them to their hearts’ content, but the skeptics are left in peace. Integrating Gemini at the system level forces it upon everyone’s attention and feels like a company pushing its own agenda rather than helping users with theirs. Speaking as an iPhone and HomePod owner, I can certainly confirm that an overeager AI assistant is deeply off-putting.

Considering the matter from a societal rather than individual perspective, broad AI integration is worrying because, as a colleague neatly phrased it, this reduces the intentionality of using the technology. Instead of hitting a roadblock, struggling for a moment to get past it, deciding to use AI, and then doing so, we are rushed from the first to the last stage. It reduces the friction in the process, and in this case friction is good because it encourages the user to consider whether AI is the right tool and whether they could accomplish the task themselves. Not to mention priming them to watch out for hallucinations (which may be particularly important with Gemini Intelligence’s unproven agents).

This isn’t exactly a startling insight, but it’s entirely absent from Google’s (and Apple’s) pronouncements on the topic, so here we go: AI is kind of bad for us. It diminishes critical thinking, spreads misinformation, puts people out of work, creates revenge porn, harms the environment, fills social media and art sites with boring slop, and pushes up the price of tech hardware. I’m not going to say we all should stop using AI, because that genie is out of the bottle. But at the very least, we should use it thoughtfully, with respect for what it costs and awareness of its limitations. Not just because that’s the first option on our phone screens.

So no, I’m not impressed by Gemini Intelligence. But at least Android won’t have those horrible transparency effects. Chin up, everyone.