Artificial Intelligence

The earliest successful and experimental AI program was invented by Christopher Strachey in 1951, in which the program could play a complete game of checkers at a reasonable speed.

In December 1955, Herbert Simon and Allen Newell developed the Logic Theorist. This first non-commercial artificial intelligence program eventually would prove 38 of the first 52 theorems in Whitehead and Russell’s Principia Mathematica.

However, it wasn’t until 1980-81 that AI became popular; most of the 1980s showed rapid growth and interest in AI, now labelled as the AI boom.

The first commercial application of AI in the 80s was an expert system known as RI (Robust Intelligence) which was used by the Digital Equipment Corporation (DEC) for configuring orders for new computer systems. By 1986 RI had become so efficient that it saved the organisation over $40 million a year.

In 1984, a robot known as RB5X, capable of learning from experience, was created. It was utilising self-learning software. The RB5X would progress from simple and random responses to finally being able to predict future events in its environment based on its analysis of its past experiences.

Virtual reality was born in 1987. That first introduced virtual reality gear like glasses, data gloves, and, much later, a complete data suit. At the end of the 1980s, computerised automation took over, marking the start of the development of AI digital programs.

Over the years, AI and robotic process automation have become more refined and efficient. Today, AI has become a necessity and not a luxury like in the early days. 

Today, AI is used to improve user experience; for example, most modern mobile phones can offer voice and text assistance, create more refined and beautiful photos, recognise objects and facial recognition, and process data faster.

On social media platforms such as TikTok, Snapchat, Facebook, Instagram, Twitter, and LinkedIn, AI can create posts for you, draft and target social adverts, automate monitoring, and power most of what you see on any given social network.

That might be why the “AI in social media” market is projected to grow from $633 million in 2018 to more than $2.1 billion by 2023, according to estimates from Markets and Markets.

AI has also spread like wildfire through the photography industry. For example, it can mimic the visual appearance of authentic images, zoom and enhance and predict scenarios.

Most modern mirrorless and reflex cameras are also equipped with pre-installed camera functions that use AI to identify the subject and location of the shot and automatically adjust camera settings to enhance the quality of the images.

For instance, AI-Autofocus allows the camera to decide whether the subject is static or moving, and will select either One-shot AF or AI Servo AF depending on what type of subject the camera senses (hence the ‘AI’ indicating artificial intelligence). This mode is automatically used in some Basic Zone shooting modes.

In contrast, most cameras also have filter modes allowing users to apply an AI-generated filter such as monochrome, black and white, antique, etc. In most cases, the resulting image will improve, but not always, so you may need to turn AI filters off.

Furthermore, AI can recognise some types of composition, such as scenes, beach, blue sky, green, portrait, etc., and intelligently apply AI auto-focus to enhance that scene.

By constantly receiving and analysing large amounts of information, AI can calculate which picture is most pleasing to our eyes. Based on this, it helps the camera adjust shooting parameters and performs editing software post-processing of the frame, applying specific AI algorithms to improve it depending on the source.

The Olympus E-M1X is one of the mirrorless cameras with AI. It makes autofocusing even more accessible, with Pro Capture allowing you to take the same “live photos” as Apple and watch long exposures as they accumulate.

AI seems to evolve every week, too; for example, smartphone cameras and some mirrorless and reflex cameras fitted with AI can analyse scenes and focus on hundreds of different objects.

Next, we have colour control in which AI can adjust the correct white balance in post process for each shot and tweak the colours to make the grass greener, the sky bluer, and the skin softer.

Regarding night and low light photography AI has it covered for you if you’re a cell phone tog. AI can asses whether the camera is in your hand or placed on a tripod, from there, it can select the shutter speed accordingly, and dependent on the subject and lighting scene, the correct exposure and number of frames will be selected pixel-by-pixel.

At the end of the process, the AI analysis all the frames that are taken, culls the unlikely ones and merges all the successful ones into one final image.

Artificial intelligence can also assist with image stabilisation by calculating the level of shake and dynamically compensating for it at any given moment. However, this type of assistance is commonly seen on smartphones.

Several camera manufacturers have already taken advantage of AI. The Fujifilm X-T4, Panasonic Lumix GH5S, Olympus OM-D E-M1 Mark III, Nikon Z7 II, Canon EOS R5, Sony A9 II and Sony’s Alpha 7R V are all powered by advanced AI, and it will be included into most camera brands on the market by 2028 if not sooner.

The OIA (Office for Artificial Intelligence) has estimated that by 2040, AI will take over other 485 million jobs worldwide, starting with customer service representatives, receptionists, accountants and bookkeepers, salespeople, taxi and truck drivers and retail services.

However, will AI replace photographic artists? The short answer is no!

AI can only work with human inputted data, anything more than that would take on more than it can handle and machines are not built that way.

So, the machine becomes useless when inputting data into the device does not include a new area of work or its algorithm does not include unforeseen circumstances.

In addition, AI lacks emotional intelligence, meaning it lacks creativity, empathy, coaching, team building, communication and influencing. No amount of human expertise can input emotional intelligence into a machine or program.

AI does not have soft skills, skills that every human worker has, for example, attention to detail, critical and creative thinking, interpersonal skills and more.

AI cannot build a rapport with new clients; AI also needs humans to operate it. AI was designed to complement humans rather than compete with them. And while AI is estimated to take over 485 million jobs worldwide by 2040, these roles will still require humans to code, power and assist AI.

AI also lacks commonsense and is incapable of reasoning and contesting facts to the degree that humans can and constantly requires fact checking by humans. AI has made much progress regarding visual and audio art over the last few years. However, it lacks the creativity of an authentic artist.

Furthermore, AI lacks expression, imagination, feelings, meditation and pragmatism. Art is also a discipline, and AI does not have that either.

As mentioned, AI can only work with human-inputted data; anything more than that would take on more than it can handle. Therefore, AI will never be able to work closely with human clients as we can.

Conclusion

AI is being used in the photography industry through image processing automation. For example, AI can automatically assist with exposure, contrast, colour balance, shutter speed, ISO, aperture, focusing, composition, framing, and filters and can help to create more visually appealing photographs.

Furthermore, AI can assist the photographer with online programs such as chatbots, booking meetings, blogging, web design, advertising and social media posts.

As with most things in life, AI also comes with many disadvantages. AI is expensive, unpredictable, unreliable, requires constant human programming and supervision, lacks creativity, makes humans lazy, is emotionless, and cannot improve, i.e., a chatbot can only do what it’s programmed to do.

There has been some talk lately in photography forums about AI taking over the stock image industry, however this is highly unlikely. As mentioned AI cannot get out of bed, travel to a location, and take photos or videos on its own, it has no empathy, creativity, and can only handle a few tasks at once.

Regarding “creative digital photography”, i.e. images made in Photoshop, that area of the stock image industry may advance; However, AI will not replace pet, portrait, or wedding photographers.

Clients looking for a photographer to take images of them or their pets want photos that illustrate their loved ones rather than manipulated images made in Photoshop.

The news and media industry also has strict protocols regarding misinformation. AI cannot create “real life events”; therefore, AI will never replace news and media photojournalists because news has to be real.

Finishing, it’s also important to note that AI is a form of assistance, meaning it can assist, enhance, create, remove and add etc. Recently, AI has hit the headlines in the photography circle regarding photographers that have used AI to mislead/cheat in photography competitions. Therefore, this area of photography will need constant monitoring with rules and T&Cs adjusted to make it fairer for all.