By pascaljappy | Monday Post
Futurists have long foretold the rise of the machine. Novelists put it into words. Film makers added apocalyptic video to the narrative. Robots crushing humans, AI taking over decision centers, you know the drill.
So much so that it’s become a future we simultaneously accept (it will come) and reject (but not just yet, this is still SciFi). If you believe the star businessmen of today (Musk, Gates and others) the threat is very real. When AI beat chess world champions 20 years ago, our reaction was “yeah, but they’ll never beat us at Go, that’s too creative a game for a machine, too human”. Now that AI has thrashed us at Go and that connected objects are starting to be a real part of our daily lives, more and more are joining the ranks that recommend caution.
Still, let’s not bother about that for now. Really, does it matter if your fridge strangles your cat, just for kicks, in January 2021 while Nest fries up your neighbour’s dog? No, what I’m on about today is far more dangerous for us photographers and it’s at work, today ! 24/7, even on this commemorative day. I’m talking about the Trophy Camera.
The what ?
From the website :
‘Trophy Camera’ is a photo camera that can only make award-winning pictures. Just take your photo and check if the camera sees your picture as award-winning.
Yeah, you read that right. Do not readjust your mind.
I mentioned the idea of AI taking photographs for us in my reaction to the Sony A9 launch. Technology is pushing us away from decision-making. Which is great when the decisions are repetitive, low-level, number crunching and unrewarding. But when we turn to machines to replace us in what is one of the most human of all human aspects, artistic creativity, all kinds of go se start to happen in my mind.
Just for information, here’s how this technical miracle is achieved :
This A.I. powered camera has been trained by all previous winning World Press Photo™s of the year. Based on the identification of labeled patterns, the camera is programmed to recognise, make and save only winning photos
Judging by the current results on the public gallery, it’s safe to say that we have a few more days of creative freedom ahead of us. And I don’t want to sound retrograde on all this. After all, I love Apple Photos because of its interface, that replaces traditional sliders with predefined effects sliders. And digital has saved us the pain of processing film. And I have a PhD in Artificial Intelligence, specifically, Machine Learning. And progress is good. But the fact that someone is actually defining the future of visual triumphs as a mashup of the past is pretty unpleasant. Or is that just me ? 😉
Still not convinced ? Check out beauty.ai, a beauty contest where the judges are AI programs trained with photographs of pretty people to detect one by themselves. Not only do all finalists look very much alike, revealing how narrow the evaluation criteria are, but training biases are painfully obvious in the fact that no coloured contestants were picked in any of the categories. AI is a fantastic tool (and beauty.ai is all about enabling robots to detect health issues just by looking at people !!) Let’s be careful about what we try to achieve through its techniques.
On a different, but not totally unrelated note, DS will soon be publishing an article about the photographic equivalent of the Maslow hierarchy of needs. This is a crowd-brainfart-storming initiative. So, if you’re interested, can you start thinking about what is essential, what is inspiring, what provides esteem and self-esteem (…) in the photographic world (I mean gear, personal style, cultural drive, social media … everything you can think of).
Finally, since so many of you seem so fond of your Smartphones as cameras, here is the ultimate strap-on gift for yours. For a more stable relationship.
Have a great day 🙂
#1321. Diwali and the Festival of Chariots
#1305. My view on generative AI in photography
#1296. Why photobooks are more precious than photo gear
#1291. Slightly Off Topic: AI and photography blogs
#1256. The Importance of Colour in the Work of Saul Leiter, William Eggleston and Joel Meyerowitz. Not.
#1126. Reframing Photography with Artificial Intelligence
Session expired
Please log in again. The login page will open in a new tab. After logging in you can close it and return to this page.
What a wonderful idea! Censor all the crappy photos, right from the start! Give all the complete idiots out there a camera which will assess what they are trying to photograph, and refuse to take photos for them, of anything less than an award winning subject!
Sounds like you want to fund them 😉 Pete, photographer turned VC ! Scary ideas out there, though, right?
I think you nailed it when you drew attention to the fact that the AI component won’t have anything creative to add to the exercise – just a coldly clinical application of some weird analysis of stuff that’s already BEEN done, even if it’s then applied to a different subject.
Some of this stuff is beginning to remind me of the time when my useless brother asked me to teach him to play chess. I used to play chess with the father of one of my school friends, who was a former State champion. After one evening with my brother, I never played chess again – we started straight after dinner, and around 11 pm I went to bed because I couldn’t bear it any longer. In one game – which got nowhere, anyway – he single handedly managed to destroy any interest I had in the game!
I refuse to allow it to happen to my love of photography 🙂 Could you please keep them muzzled & chained up !!
Back in 2004 I stumbled across a photographic discipline called Miksang (Korea for “The Good Eye”). The discipline is based on Buddhist principles of seeing in the “here and now”. An interesting, an actually, very effective methodology for developing you ability to see photographically. I was hooked. But there are caveats … colour, available light and handheld only; ie. this is the way we actually see … or that’s the theory. So I immersed myself in the new discipline, took all the courses, practiced the exercises and, in fact, did learn to SEE photographically much much better than when I started. But eventually, the underlying flaw of the concept started to surface. All my pictures started to look the same, same, same. In fact, this is the single greatest criticism of Miksang … look at the galleries and the books and the images all look the same, same, same. But them I’m a subversive, rebellious cuss … DON’T TELL ME WHAT TO DO!! I realized that Miksang was just a TOOL, and like any other tool, how you employ it is entirely up to you. Miksang helps me see what I want to capture, but I convert a lot of my work to BW (heretic). I still shoot 99.9% available light and only use a tripod when absolutely necessary, but that’s how I’ve always liked to work anyway.
So what has all this got to do with AI and photography? Well, everything!!
The camera is just a tool to capture information about the light hitting the sensor. Manufacturers may cram endless techno bumf into them because they can and get to charge us more money for their goods, but ultimately, its what you do with the information that comes out of the camera that matters. AI is just another tool available to us to create the images WE want. If you choose to rely solely on what the tools hand you, you’ll probably be a crappy photographer who’s images all look like all the other crappy images out there; and you’ll probably also be a crappy cook, bad driver, poor lover …. you get my drift. We all seem to want greatness as photographers, but few of us want, or are willing, to take the time to become GREAT at what we do. So we sell out to magic bullets, short cuts and quick fixes as a substitute for our own imagination and artistic sweat. It seems to me AI in photography falls squarely in that category, IF you allow it to dictate what your ART will be.
A final thought and I’ll shut up and go away. You may have heard of the composer George Gershwin (If you don’t know of him do look him up. Great music.). He never “studied” music in the traditional sense and always felt a little inferior to the great composers of his day. He was a great fan of Maurice Ravel and decided to go study with him. When he showed up at Ravel’s home in Paris he of course explained that he was a great fan of the Master and wanted to learn from him, etc. Ravel asked him to play for him. Gershwin sat down at the piano and played for some time, all the while Ravel sitting silently. Finally, when Gershwin stopped Ravel said “Why would you want to be a Second Rate Ravel when you are already a First Rate Gershwin?” Enough said.
What a wonderful story. And so true.
Yes, it is so much easier to look for cheap recipes than to do the hard, internal work. Not only is it easier but you also don’t risk facing failure, when you don’t put in the work.
Thanks for the reference to Miksang, a term I’d never heard about. I followed a similar practise reading books by Frederick Frank. All of these methods are excellent to pick you up from the ground and give you a direction. But, ultimately, if you don’t start to think for yourself, you’re bound to hit some sort of creativity ceiling at some point.
There’s a wonderful second punchline to the Gershwin/Ravel story. Ravels next question was “How much do you make from your music?’. Gershwin replied $100,000+/year (remember, this was the 1920s). Ravel replied “I should be taking lessons from you.”
Brilliant !!
Fascinating theory! Unfortunately, or fortunately, my 0.02$ is that this trend is unstoppable. It will happen. And, yes, pictures will all look more or less the same, but in a nice way.
This evoques the Airbus fly-by-computer system. It may not, according to naysayers, let the pilot practice some really, really wild manoeuvre that might save a plane in a desperate situation. But it prevents many ordinary mistakes, and, so says Airbus, the balance is positive. Elon Musk says the same about forthcoming auto-drive systems for cars (Tesla included). They will have some flaws in exotic circumstances. But they will save many lives in more common occurrences.
This is the age of the statistical optimum, not necessary of stellar performance…..
Time to get out my Otus and do something crazy-stupid, that no automated system would ever dream of doing. I like that, and I don’t have to show a result that would have me certified and committed….
Yes, let’s be crazy while it’s still allowed. I’m all for AI taking the lead when passenger security is at stake. But not for creativity “settings” 😉 But, there again, maybe I’ve got it all wrong. If computers beat us at Go, maybe they can beat us at photography, fair and square ! That would be sad but interesting. Not sure what would be left for us to do on this planet …
I am not worried about artificial intelligence for a numbers of reasons. It is limited to the programed rules that often do not understand balance in an image and emotional content. Is the camera going to have cattle prods to make you shift left? Also, where is the real creative artist gratification that I crated the image. Furthermore we would end up with a lot of images without real emotion and that are adequate mediocrity that the masses do anyway. I am not worried as this is my creative outlet and if I stop learning I am already dead anyway.
Exactly !
But can you imagine some of the smug social sites wif the photographs show were actually scientifically proven to be winning shots, bu necessity. That would be hilarious.
Re.: Trophy camera
Wow, to think that that tiny Raspberry Pi inside should be capable of handling all those creative decisions!
I guess it will soon also play Go…
( And it wasn’t on April 1 but on April 3-16 it was presented… Brave New Photo-World!)
Just as well we don’t all fall for the latest help-yourself-book, we’d soon become all alike…
( Just think how much simpler the camera market would become – one model would be enough!)
Perhaps this “art” example might illustrate why there are so few critical voices joining you (photographers are to busy photographing, and critics – well, see below…):
http://en.wikipedia.org/wiki/Pierre_Brassau
But, sadly, the ensuing laugh was not loud enough to humble the art critics.
EDIT:
“joining..”
I don’t, of course, mean here, but generally.
( Apologies for being unclear.)
Brilliant 😉 Can you imagine the art scene faced with trophy cameras strapped around all sorts of animals? Could be an art project in itself: “Is the artist needed any more, or will the poodle do the job ?” Can’t stop laughing.
Why not have AI captures? After all, academically trained “artists” from local universities and colleges have been using it for years to jury photography shows.
Well, there probably no reason why not, when you think about it. I can just imagine the backlash from the artistic scene. It would be interesting to compare the price of human art vs AI art.
In a way, the artist would be the person training the AI. Until the AI are able to train themselves, that is. Uh 😉
The story passed me by at the time, but Microsoft implemented an AI on Skype (I think) and within 24 hours had to be taken offline. It learnt from the conversations it was having, ad wa soon spouting racism, sexism, homophobia etc. Of course, it learnt from the best – the worst of is – so you cant really blame the machine as it was just holding up a mirror to the people talking with it. Hopefully cameras won’t go the same way just yet.
I thought some photographers already had a way to only take great photos – to buy the most expensive cameras and lenses regardless of any skill deficit?
Cheap cameras have had Gene recognition for years, and Sony have a feature to auto-compose (crop) portraits. We herald the rise of the convenience of the smart phone (which automates everything), decry the fall of the camera makers, then complain when they try to make it simple for the consumer.
The easiest way to censor bad photography would be to turn off social media or write an app to fry all smart phones…
Maybe we should take up oil painting? 😉
Meditation’s all that left. It’s an artform no one can judge, rank, imitate …
But you’re right, expensive gear IS the answer to poor photography so why bother with complex AI deployment 😉
I did read a funny interview with Fuji after the release of the X system, where someone said they had made the X Pro 1 larger than other mirror less cameras to appeal to the north American market. I’ve lost count of the number of internet discussions where it is empirical fact that the biggest and most expensive cameras and lenses take the best photographs. I’m not immune to the male collector gene or the male specification gene, and have some expensive equipment, but Id like to think I at least try to out it to good use beyond P mode. Ironically, I shot a portrait a couple.of weeks ago, and the model for his friend to take a picture with a camera phone. When I saw the resultant Instagram filtered result and compared it to my high end fine arts approach, I was left wondering how I could replicate what he had done with his phone. The cost of the equipment means nothing. I remember seeing a candid portrait of Queen Elizabeth II taken by Lord Litchfield on the stairs of Buckingham Palace with a Contax point and shoot film camera (I think a T2) and it was clear he didn’t need fancy equipment to create a great photo. It was quite humbling – although admittedly the location and access to the “model” were also major factors in the process. On Sunday I was struggling a little with an inexperienced model, trying to get a nice pose. Today, the model was excellent but the gristle in the machine – me – was letting the process down in spite of working with expensive equipment. In the end the equipment counts for nothing, it’s the opportunity (location, model, light, whatever) and the vision that counts. We all often focus too much on the equipment and not enough on the “art”. Guilty, m’laud!
Funny thing is that we all cheered at the small size of initial mirrorless cameras and now most of us are clamouring for something larger again. We’re all nuts 😉
That’s the thing with various formats and processing tools. The look you get from one is so different from the look you get with this other. And it’s really difficult to replicate. I shot a petrol station at night with my phone and with my Sony. There’s no doubt the Sony has a lot less noise in the shadows but the phone photo is a lot more striking and nothing I did in PP to bring contrast and bite to the Sony pic achieved the desired result …
I couldn’t find the QEII photo you mention but it has to be said she is an incredible model. The photographs by Annie Leibovitz, Julian Calder, Cecil Beaton, Chris Levine (stunning photograph) are really spectacular and probably a testament to her ability to strike the right pose. Maybe we should offer our services ? 😉
I actually much prefer “smaller” because of travel, but with full frame to give other image quality benefits, the system with lenses doesn’t save a great deal of size and weight in the end. I really hate how digital has made SLRs one or two sizes bigger than they used to be. An enthusiast DSLR with an APSC sensor is now the size of a pro film SLR a few years ago (yet the same electronics can be put into an NEX-5).
It is so frustrating when you cannot replicate a “look” with an obviously better camera. With phones it’s just the signal processing – tone mapping, levels and curves adjustments etc – that give them a look (all automated software post processing). But although you liked the result, if Sony put the same software in an E mount camera, we would probably complain it took away our creativity!
I can’t even remember when I saw the photo, it was many years ago. I’m fairly sure it was Lord Litchfield not Snowden, and was taken with a Contax T2 film compact. There was an event at the palace, and he photographed her as she had come down the stairs of the palace in full regalia but effectively “back stage” away from the event. I remember it just looked sumptuous.
The success Sony has had miniaturizing cameras – and advancing what goes into them – has retooled my photographic process enormously. I don’t want to go back to anything I used previously as the toolsets we are getting now facilitate choices we couldn’t get out of a camera body just a few years ago. The integration of remote controls, flash units and video accessories has redefined what photography is all about.
All ahead full!
Hmm – so a snap happy idiot who’s just bought a whatever can take a “good” shot because he’s spent a lot of money? And a professional photographer with access to whatever equipment he needs can only take lousy photos, because he didn’t spend $10 grand on a camera body and a similar amount on his lenses?
I don’t think so. A “good” photographer can pick up whatever cam is nearby and take a “good” photo with it, and I defy the wannabe’s to do the same. It’s quite wrong to define a “good” photograph by reference to the technical qualities of the camera it was taken on.
Just have a look back through the works of the REAL greats of photography, and see if THEIR masterpieces are ruined – hopeless – junk – horrible. Of course they’re not!
To take good photographs, nobody needs to have a 56MP cam with a maximum ISO of over 800,000, capable of such extreme zoom that it can produce life size images of bees and ants, as well as photographing the man on the moon.
A far better starting place is to have “the eye” – to study light in all its forms and senses – to learn about composition – and to sit back and think, “what am I doing this for? – what am I trying to achieve, to say?” No equipment – and no amount of equipment – will provide a sufficient alternative to doing this.
I guess this really opens up the whole debate of “fine art” vs “art”.
It’s hard to make a good fine art photograph without top notch gear and good technique. To an extent, imagination is not as important there as mastering a specific subject type and look.
Whereas great art (thought and emotion provoking) can be made on crappy gear.
I think the good photographer picks the gear that best complements his/her vision. We can expect a bunch of stunning action photographs from the Sony A9 users, but very few brilliant astrophotos. Gear doesn’t make a photographer but some excellent photographers are so locked into their style and gear that they fall apart without it. Kai Wong used to invite pros to shoot with crappy gear and some produced decent shots while others just didn’t 😉 It was very funny to wacth.
This sentence says it all, for me –
” . . . the good photographer picks the gear that best complements his/her vision . . .”
Thank you, Pascal – that is very well put !!!