0:00 This episode is sponsored by MPB. One of the questions I get asked the most online is people will message me and say, "Can you come across to my profile, take a look at my images, and just [music] give me some feedback?" And I get it. When we're starting out, we're all trying to work out how we're doing. So, it seems like a good idea to get hold of other photographers who we feel a bit further down the road from us so they can tell us if we're making any mistakes or give us some advice to make our work better. I suppose when we're starting out, we could go and book ourselves a professional review with another photographer, but but expensive those can be. So, it seems like the best option is to find somebody whose work we like and to try and get them to tell us how we're doing with our work. But the fact is that most of us just don't have time to be giving everybody else free reviews who asks for it. So, what do we
0:45 do? The other day I was on a call with my friend Carl, and he mentioned to me that he'd just taken one of his images and he'd actually uploaded it to AI and asked AI to give him feedback and let him know how he can improve the image. And he showed me the feedback that it gave him. And as I read through it, I thought, you know, some of this is a bit janky and doesn't really make sense, [music] but there were a lot of good points in there. And that got me thinking. I know that a lot of us are quite suspicious of AI and we have mixed feelings about how it's changing things so rapidly or even how it really works. And that's all fair and good, but with every technological change, there come opportunities for us to use that technology to our advantage. And getting feedback on our images might be one of those advantages. So, I thought that in this video, I would put this all to the test with you. I'm going to give it a few images and see what [music] it says
1:31 about my own work. And then in the second half of this video, I want to give you some advice on how to use AI to get feedback and then be very honest with you about some of the caveats and the limitations and the things to watch out for. >> [music] >> So, let's jump in. So, I'm here in the free version of Chat GPT just on my desktop. And let's try something out. Let's just pull in a random image and ask for some help with it. So, I'm just going to go here to plus, add files and more. I'm going to come up here to some selected images that I've got just to see if we can try something out. And let's try this photograph I took of a dog waiting for its owner outside a store recently. And I'm just going to ask Chat GPT. I'm going to say, "Can you tell me what you think of this
2:16 image and give me some notes for improvement?" So, this is a very, very basic prompt, right? We're not We're not doing anything fancy here. Let's just see what it comes up with. If I hit send, it's going to take a second to think. Really doesn't take very long. And this is what it gives me back. This is a lovely atmospheric photograph. There's an analysis Here's an analysis and suggestions for improvement. Overall impression, the image captures the charming and emotional moment of dog standing at It's basically telling me the image I've taken, but it's proving that it's actually seeing the reason I took the photograph. A dog standing on its hind legs peering eagerly into a softly lit doorway. The lighting, composition, and color tones all contribute to a cozy, slightly cinematic
3:03 feel. So, it understands what I'm trying to do with this image. Strengths, it says there's good mood and storytelling, warm light in the doorway. It says composition is good, the framing works well, dog is placed off-center following rule of thirds. I'm not sure I did that deliberately, but it's kind of right. And the lighting is a mix of artificial warm light coming in from the from the building and also cool ambient light giving great color depth. So, it understands kind of what I was aiming for. Okay, so here at the bottom it's got suggestions for improvement. It says one, exposure balance. The illuminated door panel is slightly overexposed losing some texture. You could reduce highlights or recover detail in post-processing. And I guess it's kind of got a point. My eye is drawn to that top right corner. I don't
3:50 think it's overexposed. I think it's wrong. I could have pulled the highlights back a little bit more, but I wanted that light source to still feel like a light source. So, I don't totally agree with it on that one. Focus and sharpness, the dog's fur is well detailed, but a touch more depth of field might keep both the dog and the door frame crisp, especially around the paws. So, it thinks I should have a deeper depth of field to include more detail in the background and the door frame. Again, I kind of like that the shallow depth of field separates the dog out from the background, but I could see that that would be an interesting image, too. I don't think it's wrong in suggesting that. I just went for a slightly different choice. Foreground cleanness, the lower part of the image has some visual clutter, the dark muddy paws for instance. While this adds realism and a
4:37 a subtle vignette or selective dodge and burn could guide the viewer's eye towards the dog's face and the lit window. In this case, I kind of agree with it, actually. I could have darkened down, especially this light at the bottom here, which draws attention. I like the muddy paws of the dog, obviously. It's half of the point to make the image. But yeah, I could have shaped maybe a bit more burning at the bottom and perhaps in this top corner at the back right. But yeah, okay. I'll take that as a point. I could have done that. Color grading, consider enhancing the cinematic feel by slightly cooling the shadows and gently warming the highlights. This would further emphasize the contrast between inside and outside. I agree with that, too. I I have already graded this, as you can see. But yeah, I could have pushed it further. Personally, I like quite a clean-looking
5:23 image, so I don't push my color grades too far, but that's not a bad suggestion. That's a That's potentially a good way to go. And then cropping, you might experiment with a tighter crop on the right side to remove some of the empty wall space and focus more attention on the dog and the light. This is something I see AI doing a lot. It It likes to have quite a tight frame on the subject in the image. So, it wants me to push right in on the dog here and have less of the window in the frame. It wants me to just focus in on the subject and push right in. I personally did this deliberately. I wanted kind of an equal balance light and dark. I wanted more of the window framing because it suggests the owner inside and what the dog's actually seeing more. It includes the the
6:09 presence of the owner that's out of the frame without having to show them. So, this is obviously a deliberate composition. I didn't do this by accident. And I want this space in the frame. I don't want to push it any tighter. It would lose something for me. So again, this point I don't take from AI. It It constantly wants me to focus in and crop in more on just the subject, but it doesn't seem to understand that having more of that space tells slightly more of the story. Let's try another one then. So, let's come up here to add photos. Let's go to this image here, which is one I took recently down in London. Now, you'll notice that a lot of this is very complimentary and it's not really pushing me very hard. So, I'm going to ask it to ramp up the level here a little bit. Can you give me
6:54 feedback on this photograph uh Funny I could spell. And don't hold back. I want a brutal a and honest opinion opinion on how this image uh comes across and giving giving me uh useful tips for improvement. And I'm going to be a bit more specific on this one.
7:40 I'm going to say uh I am a street photographer working in London and I've been photographing for about 10 years, say. So, it has an idea that, you know, I don't want beginners feedback. I want something, you know, a bit more specific. Uh and then let's say um this image will appear as part um as part of a series of images on London in my own upcoming
8:27 book of street photography. Let's see what it says to this. So, it always starts off with a compliment cuz it's AI. So, it says, "Excellent. This image has a lot going for it. And since you've asked for unfiltered critique, I'll give it to you straight from a visual and editorial standpoint. First impressions, it's a striking image, bold color contrast, confident framing, and a contemporary subject that immediately situates us in an urban modern London. The orange wall is visually dominant, almost theatrical, and the man's dark suit pops perfectly against it. It's clearly intentional and has the polish of a photographer who knows their visual language." Again, I don't totally trust this compliment. I think it's telling me that cuz I told them you know, I told it I'd been doing it for 10 years, so it wants to kind of
9:13 let me know that it's picked up on that. I don't really believe it thinks this is the polish of a of an experienced photographer, but anyway. That said, there are aspects where it can go from strong to outstanding, particularly in pacing layering. Again, the strong to outstanding, I don't necessarily think this is such a strong image. I think it's it's fine. It's good, but it could be a lot better. So, it's it's really being too kind to me still, even though I've told it to be straight with me. Strengths, we'll go through this really fast. Color and contrast, yes, obviously, reminiscent of Saul Leiter and Alex Webb. I wish that was true. Composition, the man walking dominates the frame, posture, suit, phone, all scream contemporary London professional. Yes, secondary figure in the background also on a phone adds a satisfying echo. Yeah, it's picked up on that. The lighting is punchy, harsh midday sun, but used effectively, you didn't shy
9:59 away from the shadows and the boldness pays off. The crisp shadows enhance the geometric strength of the image. Good, I'm glad it's picked that up. Context, the look painted on the road is a brilliant London detail and it situates the viewer without clichéd landmarks and the scenery, blah, blah, blah, blah. So, it's actually picked up like what I liked about this image as well. It's got this look look mirrored in these two people as well. So, it's worked those sorts of things out, which is to me, that's quite impressive. So, the weaknesses, let's see how rough it's going to be with me. It's overly clean over clean composition. It lacks grit. The image is slightly too polished and not in a good way if you're aiming for an authentic documentary street series. That's a brilliant point. It borders on fashion editorial rather than candid street photography. The cleanliness of the orange wall, the symmetry, the subjects pose all feel a bit staged even
10:46 if they're not. And if your series aims to represent London as lived, consider moments with a touch more unpredictability or tension. It's picked up something brilliant here actually cuz this is something I feel about my photography overall, that it is too clean and it's something I'm trying to undo in my photography, that it it looks too tidy. Everything looks too neat because I like I like a clean frame. So, that's a really interesting thing and it's picked up that my book that I said I wanted to do is on street photography in London and this might feel staged or too clean or not gritty enough to represent London and I think it's absolutely right. Narrative thinness, there's visual tension but not emotional sociological tension. The photograph is beautiful but not probing. It tells us about aesthetics
11:31 not much about the people or the place, businessman in the city. You could lean into irony or juxtaposition more, maybe catch an emotional flicker, something that undercuts the city's surface tension. Again, I think this is exactly right. There is no narrative really in this image. There's no emotionality and there's not enough of a story in it. It's aesthetically interesting and very very clean and represents that kind of clean contemporary modern London, but there isn't much emotional narrative in it. Spatial flatness, though the wall is powerful, it compresses the depth of the frame. The car's presence add a bit of layering, but it's truncated and the right side feels underused. The composition could breathe more if there was a stronger middle ground. Again, this is exactly right. Like I can see that and and this happens with a lot of my photography. I do like that clean
12:17 very simple layered frame, but it's saying maybe you want to think about more depth in the frame. And this is something I've been thinking about my photography in general, so it's not about point. Color over substance is dangerous balance. Color dominates to the point where it risks being the subject. In a book context, if too many images lean on strong color blocks without emotional complexity, it might start to feel repetitive. Make sure your sequencing alternates between color driven and story driven frames. It's taking into account the fact that I'm talking about wanting to put this in a book and it's warning me of the dangers of just making images like this, which if they feature too much, they might start to dominate and take away from the the flow of the book overall and I think that's also a pretty good point. Suggestions going forward, it's got experiment with a wider frame. It
13:02 obviously wants me to bring in more context around the subject potentially. Not a bad suggestion, I could try that. And this is obviously about images going forward rather than fixing this individual image. Then looking for visual irony, that look on the ground, trying to work out a way to frame it maybe to to play into the scene more. Shadow play, since you handle light well, try using shadow as a compositional narrative tool. It's something I do in my other images, but yes, maybe if I came back at a different time of day to this particular scene, there might have been an opportunity to use more shadow on that wall if a building cast shadow on that wall. Edit ruthlessly. For a book, this image could work best if paired or sequenced cleverly. Maybe think of the opposite, something chaotic or messy. That's an
13:47 interesting point trying something like that. Tonal consistency, decide if your book will embrace this high contrast saturated or balance it with softer atmospheric images. Yep, I think that's also a good point. How will this image fit in? Then the verdict, this is a technically excellent image, color savvy, confident, clearly shot by someone with control and intent, but it's a bit too safe, missing a degree of unpredictability or rawness that makes truly great street photography cut deeper than surface level design. If I saw this in a book, I'd think this photographer knows their craft. Now I want to see what they feel about London, which is hilarious that AI seems to be quoting itself, but what a great point, you know, it's saying to me this might be too clinical. It might be too neat and I feel that way about my photography
14:34 in general. So, in one image, it's uncovered something that I pick up about my own work already. Before I jump in and give you some practical advice on how to use AI to get feedback for your images, I just want to thank the sponsor of this video, MPB. If you're like me, you're always working to perfect your kit to just get the stuff that you really need to do the job well and that might mean getting that next camera body or camera lens that's going to take your work forward or it might mean selling off some of your equipment so you can downsize and make that kit even more portable. Either way, as we're approaching the end of the year and getting ready for next, now is the right time to change your gear. Selling with MPB is super easy. I've done it myself a number of times. You just head over to their website for your quote and if you're happy with what they're offering you, you accept the quote, pack up your item, print off a shipping label that
15:20 you put on the front of the box and then a delivery driver will come to your door with an 1-hour slot to pick up that package and all the postage is free. You don't even need to go to the post office. And then MPB will receive your item, they will test and check it and if everything is what you say it is, the money will be in your account soon afterwards. And buying used cameras or lenses from MPB is just as simple. We all know that sometimes buying used equipment online can feel a little scary in those online marketplaces cuz we don't really know who we're buying from or what sort of condition the equipment will arrive in. But when you buy with MPB, every item you purchase has been individually tested and checked and comes with a free 12-month warranty, which personally gives me a lot of peace of mind. And they also have over 40,000 reviews on Trustpilot with an excellent rating, which again just makes me feel safe in the purchase that I'm making.
16:07 So, whether you're looking to buy or sell used camera equipment, MPB are a fantastic option. I will leave a link at the top of the description below. Go along and check them out. So, let's get back to it and I'll give you some practical advice on how to use AI to get feedback on your images, some of the things I think that I've found have worked and also let's be honest about some of the limitations around this new technology. A word of warning before you get started, you have to consider whether you even want to upload your images to AI or not because we don't really know how these models work and there is every chance that that image is going to be stored in that AI model and may even be used to generate images for other people at some stage. So, we don't want to be naive about that. The way I think about it personally is if I upload my images anywhere on the internet, there's a good chance that it's going to
16:53 be scraped by AI at some point. So, I'm not avoiding all of this by just not directly uploading my images to AI. And at the end of the day, I still own the rights to all my images. So, if anything does come up somewhere, I can pursue that legally. But I don't want to sit here and pretend that there are no risks. Then I want to say that I think AI is very helpful for beginner photographers, those who are just starting out. But for those of us who've been at this for a little while, I think it's less helpful because AI is great at critiquing the technical side of photography, right? So, it's good at looking at lighting and composition and color balance and dynamic range and giving you some feedback on that, which is very helpful stuff if you're a beginner photographer. But for those of us who've been doing this for a while, we probably know most of the technical tricks there are to know and now we're choosing to break some of those
17:38 technical rules deliberately to get a voice to our work, to to give it a flavor that's all our own. If at the moment all AI can really do is give us that technical feedback on our work, but it doesn't yet understand artistic expression, then it can be helpful for beginners, but it's not going to be helpful for those of us who've moved beyond the technical considerations of our work. I'll give you an example. I don't think AI really understands negative space because I've shown it some of my images where I'm deliberately leaving big blocks of the image in deep dark shadow and it it wants me to zoom in. Like it did with the picture of the dog, it wants me to crop in on the subject and and lose that what it thinks is just empty space or a mistake. It doesn't understand what it's looking at, but when I show that to a human being, they immediately understand what I'm
18:23 doing. AI just doesn't yet. So, if you're in the first few years of your photography journey and you're still trying to get to grips with the techniques, AI might be quite helpful. But for those of us who moved on a little bit and we've defined a style to our work, AI probably has less to offer us at this point. But the good news is we're probably not the ones who are reaching out to strangers asking them for feedback on our work. So, AI actually fits a good little gap at the moment. Linked to all this, I would also say stand by your artistic intentions around your work because like I said, AI is good at techniques, but it's not great at understanding art. It doesn't feel, so it can't anticipate what we might feel when we look at an image and we take it in. A human being can look at a piece of art and we can decide how it moves us, what it makes us think about,
19:09 what it makes us feel, but AI just can't do that. And we might think to ourselves, well, maybe we could describe it, but I think even if you tell AI what your artistic intentions or even your inspirations are, all it can do is give you the middle of the road feedback on that sort of work. And we know that when it comes to art, that's not good enough. AI can't tell if a photographer has intentionally broken the rules by say darkening down their exposure to create a lot of shadow in the image to make it more dramatic or maybe that they've decided to throw off their composition and put things towards the edge of the frame because they know that creates an anxiety in their viewers and that's the message they want to put across. We might have done that deliberately, AI might think that's just a mistake. To give you a ridiculous example, I uploaded this image of a bathtub in a
19:55 landscape, which I took a few years ago, and asked AI for feedback. And it said to me that it thought that this would be a much better landscape image without the distracting bathtub in it, which is, of course, true and ridiculous at the same time. A human being would have looked at that photograph and understood immediately why that image was interesting, but it took a lot of explaining to AI before it would accept that image for what it was. I think it's very normal for most photographers that we start our journey in the techniques, and we learn lots of tips and tricks, but at some point in the middle of our journey, we start to work out what we want to say with our work. And we often leave a lot of those techniques behind because they were fun to learn, but they're not useful now. And we use a few techniques, and we break the rules because what we're saying with our images is more important than how we're making them. AI is useful for that first
20:41 half, but it's not very helpful in that second half yet. One of the most helpful pieces of advice I can give you is treat it like a conversation. I think so many of us, for so many years, are used to interacting with something like the Google search bar, where we just put in a question and get a single answer out. But this is not that. AI is different, and you often won't get something that good on the first pass. But the thing is, you can go back, and it remembers what you've said before, and it will build on itself and give you more and more useful things. So, if we talk back to it and try and refine it down, we'll get more helpful answers as we continue that conversation. And be specific, because the more information you give it, the better results that you'll get back. So, tell it what you want feedback on specifically in that image. Explain what you're trying to say or do with
21:27 this particular image, why you were drawn to this scene, perhaps, how you shot it, and maybe if this is a standalone image or if this image is going to fit in a series somewhere, because you're going to get better feedback. It's also worth mentioning that Carl and a friend of his are actually working on an app to help photographers interact with AI to get feedback on their work, where there are easy prompts and buttons to click that let it know what kind of category the image is, what your level of experience is, and how harsh you actually want that criticism to be that comes back to you. So, I'll leave a link below in the description if you want to go along and check that out. Then lastly, it's important to say, take it all with a pinch of salt. AI can sound very authoritative, even when it's completely [music] wrong. So, you're still going to need to use your own brain, work out which information and feedback is
22:12 [music] actually helpful to you, and use your discernment like you have to on any other corner of the internet [music] to work out what's true and what isn't, and what's helpful and what's not. You will also have noticed that AI is incredibly complimentary, which feels great to our ego, right? But we don't just want to sit there and just take compliments and AI told me I'm a great photographer. We want to push ourselves, be better all the time, which means we have to stay honest about where we're actually at in our journey. We have to ask it for harsher criticism, [music] so we can find the growing edges in our work, because we're not just trying to build up our self-confidence, we're trying to actively get better at our photography one step at a time. So, I hope you found some of this helpful. We really do live very exciting time technologically, and >> [music] >> like you, I share your concerns about how fast things are changing and what's really going on with AI. This is
22:58 definitely a mixed bag, and I don't want to take away from that. But like I [music] said, with any technological advancement, there will be ways we can use it to our advantage, and this might be one of them. Does any of this replace our need for human interaction, or getting professional opinions, or actually finding a real-life mentor? Of course not. Those things will always be infinitely more valuable to us than this technology is. But in the meantime, we do get to use this technology to fill in the gaps. And personally, it's something I wish I had access to when I was starting out. So, if you are that beginner photographer in that exciting first half of your journey, where you're learning those techniques and trying to improve, for the first time in human history, you have access to a technology that can [music] give you customized and personalized advice on your work and suggestions for improvement. And if any
23:45 of you have already been using AI to get feedback on your images, and you have any advice or tips or things you've used or things you've tried that have been helpful, please share them in the comments below. This is new technology for all of us, and any [music] advice, I'm sure we'd all appreciate. Thank you so much for watching, and just before I go, if you want a way to support this channel [music] practically, the best way you can do that is picking up a copy of my Parable Magazine. There are now six editions available with more on the way, and each one features writing and photography around a theme. You can pick up a copy like this in physical format in A4 size on beautiful uncoated matte paper, or you can download a digital version, which is a PDF [music] and features an MP3 of me reading the magazine to you. I'll leave a link down in the description, [music] so you can go along and pick up your copy.
24:42 >> [music] [music]
