AI and Photography

AI Generated Image

The featured image of this blog was AI generated. I used Google to identify a free online application, entered a few words and seconds later an image was generated that I could download. With a paid subscription I imagine that the quality would improve dramatically. In the past few weeks I have had numerous discussions with people about AI and its impact on the creative sector. I have even recreated a client video using AI which would have saved the cost of a studio and a presenter – the results aren’t nearly as good but for a piece of fictional content I can see clients wanting to explore AI as an option.

AI and Photography
GQ Kate Winslet Retouching

A new technology for an old trick

There is a lot of furore about AI and photography, particularly when AI generated images are winning prestigious photography prizes as happened with this year’s Sony World Photography Award, but image manipulation is nothing new.

In 1934 a photo attributed to Robert Wilson transpired to be the wooden head of a dinosaur stuck onto a plastic toy submarine and the ensuing photo was claimed to be the famous Loch Ness monster. Magazines have been under constant pressure for the manipulation of the images of models and celebrities that adorn their front covers. It’s known as ‘retouching’ but as Kate Winslet found out when she appeared on the cover of GQ magazine, it can also result in losing several pounds and growing a few inches taller. These days celebrities embrace their blemishes and are only too happy to share their ‘no makeup’ selfies on social media.

The destructive power of AI and Photography

Earlier this year a photograph (below) appeared on social media that appeared to show black smoke rising up within the grounds of the Pentagon in the US. As the image spread across communication channels so did concern that the US was under attack and that was immediately felt in the markets as investors sold and share prices fell. Fortunately, the image was quickly verified as a fake and normal service resumed but this demonstrates the potentially destructive power of an AI image that can be created in minutes using online applications.

AI generated image of attack on pentagon

AI and fiction

The use of AI generated images as described above is plain deception or, if we are feeling generous, it is a work of fiction. As a work of fiction we should have no problem with AI images, afterall we are comfortable reading fiction and being entertained by fiction on our TV screens – none of really belives that Superman can fly, but we are willing to suspend disbelief and enjoy the special effects created for our enjoyment. (The video I recreated using AI was presented as a work of fiction, as was the original version)

The concern is that AI images might be used to portray facts or a truth about a situation or an event. This is a frightening concept at a geopolitical level but is also worrying at a corporate level. One contributor to LinkedIn has written about how he used AI photography to place his products in real-life scenarios as a way of selling more units – this is known as the Social Proof principle defined by Robert Cialdini as a way of influencing decision making – and we may choose to regard this as harmless marketing but how far are we willing to be deceived and how will we know if the line has been crossed?

AI and Documentary Photography

One of the drivers of documentary photography is bringing an aspect of the world to an audience that might otherwise not be able to see it, either because of geography or restrictive access. Businesses will often commission photography for the same reasons – they may want their shareholders or stakeholders to see what great work they are doing in some part of the world, but how would we feel if we found out the images being shown to us as ‘evidence’ of good work were in fact generated by an App?

It is not only businesses who can use AI for their own purposes, earlier this year Amnesty International came under fire for circulating an AI generated image (below) that claimed to show a young woman being hauled away by riot officers during a protest against the state. The image was widely circulated and it was only an error in the colours of a flag that alerted people to its real origins.

With documentary photography truth is always about perspective but even to convey a particular perspective you still have to be able to photograph it. If I want to portray a street as deprived and run-down I need to make photographs that support the story, if they aren’t there I can’t make them up. Except that now I can, now I can photograph a street and use AI technology to place whatever I want in the street that supports my story. Which is fine, if I am creating a work of fiction but not if I want to manipulate the audience into believing something as true.

We aren’t going to roll-back on AI, it’s here and it will be used. It will be down to individuals to use it with integrity and it will be up to the public to insist that it is not used to manipulate them, and if they find that it is, then they need to demonstrate their displeasure.

In the same way that AI brought Amnesty International into the headlines for all the wrong reasons, other organisations will need to be wary of their own reputational risk if they condone the use of AI images for anything other than fiction.

AI and the Creative Sector

In the same way that reduction in the cost of technology and the increasingly sophisticated ‘auto’ settings on devices has eroded the value of the creative sector by confusing technical capability with creativity, AI will do the same I’m sure.

The video I recreated using AI, originally required a camera operator (who also did sound and lights – both were separate roles when I first started in the industry), a presenter, a studio with a green-screen set-up and an editor to composite, check levels and output a clip. To do this with AI I picked an actor from a menu, picked a voice, chose a news channel background and uploaded the script. Five minutes later I had a link to a downloadable file. It doesn’t take a genius to consider the impact this will have on technicians, studios and presenters/actors over time.

However, AI photography or video cannot create real case studies, cannot report truth, and is no substitute for genuine creative collaboration. How many times has a photographer or film director been onsite with a plan that has changed in response to input from others? I know it happens all the time in my work and that the interaction almost always leads to a better outcome that cannot be achieved with one person sitting in front of a screen.

I think AI will negatively impact the creative sector, I think artistic street photography in particular will be upended by AI images, but I also think that clients and the public will quickly tire of such imagery and demand reality.