Search
  • ami Creates

Artificial Intelligence & Photography: The death of a craft?


When realist French painter Paul Delaroche first saw a photograph in the 1840’s, he exclaimed “From today, painting is dead!”




Every few years, the arts and technology collide to create something new. At the time of Delaroche, painting was considered the highest form of visual art and the only way to depict reality. His reaction to photography reflects artists’ timeless and cyclic anxieties associated with technological progress. However, progress and new technologies are almost certainly not a bad thing, although scepticism is not unusual in the early years of adoption.


A more recent example of this is the popularisation and shift towards digital photography, as digital camera sales surpassed film camera sales circa 2003. Many photographers opposed it at the time and some still do. Some argue that digital photography lacks physicality, as the chemistry in analogue photography is discarded in favour of a computer interpreting light with the help of 0s and 1s. For others, this is synonymous to a rupture with the natural and human nature of photography as computers leave little to no space for imperfections and seemingly lack emotional intelligence. AI and machine learning is the most recent technological advance in photography and their adoption remains as controversial as ever.


At ami Creates, we firmly believe in innovation, both from a technical and creative standpoint. AI already plays a part in the services we offer, and we are exploring new ways in which this exciting technology can take our craft to new levels.





Artificial Intelligence vs Machine Learning

Artificial intelligence is a field of computer science and a technology which can be summarised as the quest to create computers which can learn, reason and act intelligently. Machine learning, as a subbranch of AI, was first coined in the 1950s by one of its pioneers, Arthur Samuel, who defined it as “the field of study that gives computers the ability to learn without explicitly being programmed.” Nowadays, the two terms are often used interchangeably as the most recent advancements in AI predominantly pertain to machine learning.


Traditionally, programming involves giving a computer a clear set of instructions for it to accomplish a specific set of tasks. Although this approach is ideal for clearly defined tasks, the more extensive the task, the more complicated and tedious it becomes to write clear and concise instructions for the computer to follow. This is where AI and machine learning shine and revolutionise the way in which computers interpret information. The concept is quite straightforward, it involves letting the computers learn how to program themselves through experience, as opposed to having a human guiding them through each step.

This is comparable to how humans teach one another; it would be impossible to teach one person everything there is to know about every single subject. As humans, we are given the tools and skills to learn so that we can then go on and apply said skills to specific areas of interest.


The process of machine learning therefore heavily relies on data, which comes in many shapes and sizes, such as numbers, photos or text. This information is fed to computers and is then used to train them. The more extensive the data set, the more efficient the machine becomes at accomplishing a task. Through learning, the computer acquires a form of intelligence through which it can recognise patterns and subsequently predict them.





Photography & AI

At the time of writing, digital photography has become the norm. Alongside this more widespread acceptance of the digital, AI has slowly but steadily seeped into photography and technology. Google, Apple and Adobe have all been at the forefront of research in AI and are responsible for its implementation in the realm of photography. Since the early 2010s, these companies, amongst others, have sought to teach machines how recognise scenes, people and objects. From smartphones to professional-grade cameras, most if not all modern cameras now ship with AI that we all use daily, sometimes without even realising it.


There are an innumerable number of use cases for AI in our smartphones alone. Facial recognition, for example, is the technology which allows us to unlock our phones with just a quick glance at our front-facing camera. It successfully uses AI to recognise and read people’s facial features as it was trained extensively, using millions of human faces as data sets. Portrait mode on iPhones also uses AI to separate a subject from its background by simulating a shallow depth of field. Additionally, every smartphone has a post-processing AI which will read and interpret the scene captured and improve it in a variety of ways. If the scene is too dark, the AI will brighten it. If the scene is too blurry, the AI will rectify that by adding some contrast and so on and so forth. This is all possible thanks to the thorough training that the AI has been put through, as it learns what constitutes a good or bad photograph.


Moreover, AI is also used extensively in photographic software. The AI tools and functions found in dedicated applications are very similar to the ones outlined above, albeit offering more control to the user. Whereas this process is automated on our smartphones, software such as Adobe Photoshop allow more freedom of choice to the human using it and thus offers more creative control. This is a big part of what we do at ami Creates, blending cutting-edge AI technologies alongside human expertise and experience, therefore combining the best of both worlds.



What's Next?

If Delaroche was a photographer today, he would undoubtedly say that photography is dead. The shift to AI in photography is far from perfect as it is still in its very early stages. Over the next few decades, the technology will continue improving, as the industry will grow to accept it as an extension of the photographer rather than a substitute. Inevitably, computers still exist as nothing more than machines whose sole purpose is to process numbers. Although machine learning offers some semblance of intelligence, it will take more than a lifetime before any form of emotional intelligence can be taught and reproduced. Artistic qualities and sensibilities found in photography will therefore remain as a human prerogative for the foreseeable future, with AI serving as a tool to help realise and express these visions.



Discover more about our bespoke creation services at www.ami-creates.com