Alphabet Inc’s Google is racing to stuff its products with the most advanced artificial intelligence features, including some that will make you question everything you see and hear online. Its new Pixel phones make it easy to manipulate photos, adding people who weren’t in the original shot or moving their positions. You’ll be able to record phone calls, albeit with a disclosure to the person on the other line, and get a detailed summary of the conversation. These cool, if creepy, features point to a disconcerting direction that AI tools are taking us in as they get built into more phones. The easier it is to manipulate the content we capture on our devices, the harder it’ll be to trust what we see on them too.
For the new Pixel 9 phones, which go on sale on August 22, Google has concentrated most of its AI prowess — abetted by the powerful G4 chip inside the phone — on the camera. The Add Me feature is billed as an alternative to the awkward angles and poses of the selfie. You take a photo of a friend, then get them to take a photo of you. Google’s AI stitches the two so it looks like you were standing together.
In the age of social media, where the statistical probability of looking at an untouched photo has diminished considerably, isn’t that going to make it even harder to determine what’s real? Not at all, according to Google’s Rick Osterloh, who took charge of Google’s Android platform earlier this year. In an interview with The Wall Street Journal, he said Google was simply allowing people to edit their real-life moments, and “store the memory how they want” — no different than Photoshop, he added.
It is in fact, very different. The vast majority of people who take photos with their phones don’t pay $23 a month to use Adobe Photoshop. But millions of people who are likely to buy new AI-enabled phones from Apple Inc., Samsung Electronics Co. and Google will be able to manipulate photos with a few taps. Apple’s forthcoming iPhones will have a Clean Up tool to remove objects and people from photos. Samsung will let you move someone in a photo so it looks like they’re facing someone else.
These features are marketed as conveniences, but they’ll also make us more likely to question the accuracy and reality of photos far more than we do now. In his interview, Mr Osterloh also defended a TV ad promoting Google’s AI tool Gemini during the Olympics, in which a father encourages his young daughter to use AI to write a letter to an athlete to tell her “how inspiring she is.” Google was criticised for being tone deaf to the real reason parents help their kids write fan mail: The process of expressing gratitude. But according to Osterloh, this is just like the transition people made from sending hand-written thank you notes to emails. “This is a way to touch more people,” he said.
Of course, the flip side of connecting with more people is that those people won’t know if they’re being “touched” by a human being at all.
Until now, the price of tech’s latest conveniences has been money, personal data and attention. Use Google’s services and you’ll be subject to an online auction for your eyeballs that shares your personal data — location, browsing history, videos you’ve watched and more — with an array of different advertising networks for targeting. There’s less of that on an Apple device, but you’re still likely spending hours scrolling on one of the most addictive tools in human history if you have an iPhone.
The price we pay for tech’s whizzy features seems subtle and abstract at first, but over time become clearer as they go mainstream. People around the world now spend about six hours a day on smartphones, often at the expense of sleep, kids or more fulfilling activities. More than 70 per cent of companies in the US now collect personal data, according to Statista, and two-thirds of consumers around the world feel tech companies have too much control over those details, according to YouGov research.
What will the price be for a broader rollout of generative AI tools? Trust seems high up there, and not just of technology companies but increasingly of what we see online, including things that are real. When Donald Trump claimed that official photos of crowds at Kamala Harris rallies were AI generated, the rumour — which was false — spread to millions on X and other social platforms because of its plausibility. Like it or not, the growing ubiquity of AI phones will force us to become more wary of what we see and hear—unless we want manipulated versions of reality to be the new normal.
First Published: Aug 16 2024 | 11:52 PM IST