Google’s Pixel 9 Pro May Have Finally Made AI Useful
We have been extremely skeptical of AI tools, especially as they relate to photography. But today during the launch of the Pixel 9 Pro, Google said its latest round of Gemini-powered AI would be actually helpful and after seeing some of what it can do, we might agree.
When Samsung launched its Galaxy S24 devices earlier this year, we were not particularly impressed with how much it relied on AI but were more let down by how poor the performance of that AI was. It was unimpressive to the degree that we were certain most consumers would try it once, see the shoddy results, and never try it again.
Today, Google showed a few new AI tools that will launch on the Pixel 9 series that directly affect photography which we think are not only useful but they are, dare we say, actually fun.
The “Add Me” tool (shown in the video above) is designed to eliminate the awkward arm-length selfie by letting two people take a group photo together in two stages. You no longer have to ask a stranger to nab a picture, you can just each take a photo in the same spot and hand the Pixel off to each other. The tech behind Add Me relies on technology similar to Google’s existing “Best Take,” which selects the best version of each portrait subject from a series of photos, combines them into a single frame, and applies it to a whole person. It’s smart and — more importantly — it works very well (at least from what we saw in our brief hands-on session).
Although we didn’t have any opportunity to test out the new Night Sight capable Panorama mode, the results we saw at the event show great promise. Google has always had excellent low-light capabilities but this advantage can now be applied to panoramic images which unlock the potential for some truly breathtaking landscapes and cityscapes at night.
The ‘What is a Photo?’ Argument Will Continue
Google also showed off a new feature called Magic Editor (best illustrated in the video above), which lets you combine generative AI with an existing photo. It’s like a smarter version of Generative Fill or Content Aware Fill in Photoshop but it works far better than both. It combines “reimagine” text prompts that leverage Gemini to create new assets with an existing image, allowing you to re-frame, change the surroundings, or add new things to a photo. The best part about this is that it works really well — it’s vastly superior to anything we’ve seen that has tried to do similar from either Adobe or Samsung. It’s immediately usable and, honestly, fun. We also found the speed with which it processes images to be much faster than the previous experiences, and subsequently far more enjoyable to experiment with.
Magic Editor will no doubt be extremely controversial — as it should. We didn’t notice any kind of indication on-screen that an image was AI generated and the results that it outputs are very realistic; you would absolutely be able to pass them off as real to many average people. Magic Editor isn’t going to make arguments about what is real and what isn’t any less antagonistic.
Google has put restrictions on what text prompts Gemini understands but users will no doubt find ways around these digital walls. All this said, when it’s used within the parameters of what Google envisions, it’s the first time we’ve seen an AI editing tool do a reasonably good job. It pulls you in and makes you want to engage with it, which is more than we’ve been able to say about similar technologies so far.
In February, a Samsung executive famously said that “there is no real picture,” which is of course, nonsense. That said, the argument of what is and is not a photo will only be fueled by what Google has created here.
It’s Not Just AI
AI was much of the focus, but Google also made some hardware upgrades on the Pixel 9 Pro and Pro XL that are worth calling out. One change that pleasantly surprised us right out of the gate was the identical camera array and performance specs from the Pixel 9 Pro and Pro XL. Other than the size difference with the Pro having a 6.8-inch screen and the XL having an 8-inch screen, the two options are identical and this is a far more agreeable approach to allow the user to simply pick what size of phone they want without losing any desirable features.
The screens themselves are also much improved. The new Google Actua displays provide a peak maximum brightness of 3,000 nits making them easier than ever to use in bright sunlight. Even the 2,000 nits maximum brightness for HDR content playback is very impressive.
Last year, we voted the Google Pixel 8 Pro to have the best overall cameras for smartphone-based photography, and yet Google surprised us with some meaningful hardware changes on the 9 Pro which have a strong chance of maintaining that position.
The main camera stays the same as last year’s model with a capable 48-megapixel camera with a bright f/1.68 aperture. While an upgrade would have been nice, it’s probably fine since this is an excellent camera overall. What did get an upgrade was the ultra-wide camera. Smartphones need as much light as possible and that 48-megapixel camera has a brighter f/1.7 aperture rather than the f/1.95 from last year. The 5x telephoto has the same light-gathering specs as the old lens but it now focuses faster and more accurately thanks to a new dual-pixel autofocus array.
Perhaps the most dramatic change is the new front-facing selfie camera which sees a big jump in detail up to 42 megapixels. The lens is also slightly wider with a 103-degree field of view. It also has autofocus. These are meaningful upgrades and selfies shot with the Pixel 9 Pro will be far more impressive than what we have seen before and doesn’t represent the major compromise in image quality that it did before.
There is a lot still to test with the Pixel 9 Pro and we’ll go into detail in our coming full review. For now, though, we remain optimistic that Google may have made the most complete AI-powered handheld yet thanks to the actually useful AI that is melded with solid hardware.
Additional reporting by Chris Niccolls
