All this hype and scaremongering about AI in photography is missing one crucial point. No matter how good it gets, far from taking over the world of photography, AI will always fail at one critical point. Our recent history tells us why.
Sitting at my desk, suffering from the latest COVID variant, I talk to the smart device embedded in my chest: “Sirexa. Display a photograph of sunrise behind the Coquet Island as it will look tomorrow morning.”
Not that long ago, it would have dispatched a drone to get the shots from precisely the right spot, but that is rarely necessary these days. The computer analyses the Meteorological Office data and then studies thousands of photographs already on the Internet of the island and its lighthouse. It then creates predictions of what the scene looks like.
My office wall displays one of the photos. I ask the computer to animate it. The waves start lapping on the shore, the light flashes, and terns dive into the sea. “Sirexa. Add sound.”
That may have once seemed like science fiction. But every day, it is coming closer to reality. We read about how this technology is the end of photography. But maybe it is not as scary as we might suppose. There is one critical point that the scaremongers forget to mention.
I got a surprise yesterday. I was wandering through the never-ending Internet forest, and I came across something I thought was long gone. It was the virtual reality world of Second Life. Out of curiosity, I opened an account back around 2004. However, I quickly got bored with it, getting on with real life instead. Reality has been far more exciting and rewarding than any artificial world inhabited by fake people could be. Such worlds have become the basis of many computer games where non-player characters, computer-generated people that are part of the game’s story, live out their limited lives. Limited, that is, in both depth of character and longevity. People don’t care about those automatons. When they get eaten by a dinosaur, zapped by a laser, or chopped in half by an orc, the player does not grieve over them.
Second Life and other virtual worlds were supposed to be the next big thing. They allowed total anonymity so that a user could become an entirely different person, hiding behind their avatar. They were born at a time when the Internet was still very new in most people’s lives and there was a great deal of paranoid suspicion about sharing one’s identity online. As internet forums started to grow, they too encouraged users not to use their real names.
There was a bad side to this. Trolls were lurking behind false identities. They took great joy in “doxing” people, especially women – trolls are invariably misogynistic – working out their true identities and revealing that to the world. Nevertheless, right from the start, I used my own name online. My reason being was that it appeared in the telephone directory. As I was not called John Smith or Peter Jones, I was very easy to find. One or two trolls targeted me, but I was found by far more good people than the sad, hate-filled individuals sitting in their sweaty bedrooms.
Then, along came Facebook. Unlike Second Life and internet forums, it insisted on people using their real identities. This was a good thing in some ways, as it went some way to stamping out the anonymous trolling and gave some protection – albeit very little – against anonymous copyright theft. Those that did hide behind a false mask with malicious intent were exposed and sometimes sued for their bullying behavior. However, Facebook had control of our personal data that we posted on different websites, and that data was worth money. Anonymity would have been contrary to their business model.
Facebook is not alone in this, of course. When you go into a shop and they ask you for your email address or if you use a loyalty card, the shops are collecting your data so they can target you with advertising and spot trends in the markets.
If these super-rich people and big corporations see something that they think will improve their data collection, increase their business, or enhance their reputation, then they either acquire it or duplicate it. For example, Facebook bought Instagram, Donald Trump created his own social media platform, and Twitter changed the way Tweets are presented allowing bigger tweets that could be edited.
It’s also how the monetization of our photos by third parties happened. Each time you post an image online and people like or comment on it, that data gets examined and reviewed by computers and is used as part of the big data to improve company profits. Those robots look at what time the photo was posted, what hashtags you used, and whom you tagged in the post. But they will also use the data to target you with marketing. Those hashtags and the location data make it easy for the robots to identify what is in the photo, where you have been, and what your interests are. Consequently, they will tweak their algorithms to tailor the advertisements most likely to appeal to you.
No wonder that Instagram doesn't seem to actively discourage those photo amalgamation accounts that share your photos without your permission. They generate a lot of likes and comments, all adding to their collected data.
If you don’t like it, you can move to another platform such as Vero. But if that becomes a huge success, then you can be sure that will be bought by a big business too. It’s not just Facebook that will be doing this, and it’s not just done for financial ends. If you have signed up for a political platform, you can be sure that your data will be used in a way to ensure your vote.
Some people imagine that Mark Zuckerberg, Larry Page, Tony Bezos, Donald Trump, Elon Musk, and other super-rich folk are all poring over one’s individual data, discovering that you like Kellogg’s Cornflakes for breakfast and Starbucks coffee at 11:00 am on a Wednesday. But, let’s be honest, you really are not that interesting. Nor am I. It is just a computer churning your data while it compares millions of others’ data too. It doesn't know you. You are just a tiny, individually uninteresting speck in an ocean of data. The computers know a lot about you and your buying habits, as they do about the 61% of humanity that is online. But the people who own them are not concerned with which brand of camera you own, what your inside leg measurement was, or where you were last Thursday. But those computers know more about you and your family than you think. Ten years ago, there was a much-shared story about a supermarket chain that started sending baby product offers to a teenage girl based on her buying habits. This enraged her father, who complained to the store. He later phoned to apologize because the store’s computer program had got it right; she was pregnant. In the last 10 years, the amount of data has exploded and the algorithms that handle are now far more sophisticated.
Like that story of the teenage girl, we find that behavior by computers creepy. We are freaked out when we mention in a phone conversation about the latest lens, and it then starts to appear in adverts. We feel like our personal life has been invaded. There’s something unnatural and disturbing about it. Similarly, animation filmmakers know this and in movies, characters are made to look less photorealistic. Artificial humans are getting better, but they are still lacking that spark of humanity and, consequently, seem unreal.
Just as we created artificial intelligence, AI now creates art. But, a big part of the value of art is that it is an interpretation of what the artist is thinking and feeling. A machine does not think or have emotions. AI images, however realistic they may be, lack humanity. They did not feel what you feel when you held that viewfinder to your eye.
And that’s why art in any form, including photography, AI won’t replace humans. When we look at a van Gogh painting or an Ansel Adams photograph, part of the meaning is the personality, the soul behind the creation. AI creations have no humanity. Yet.