By Christodoulos Pouroutidis
When Siri was first implemented in iPhones in 2011, the service was lauded as the most capable and modern personal assistance anyone could have, and housed within a device the size of a bar of soap that was the iPhone 4S. Siri, to this day, can tell you weather forecasts, set alarms, book restaurants, and even tell you a joke – it’s able to listen to commands and act on them dutifully and accurately. What it cannot do, just like its competitors, is synthesise data to give you an accurate, detailed response to a multi-layered question. If you ask Siri “why did the French revolution happen?”, it will struggle for a few seconds until it generates a Google search of important articles that give you an insightful account written by a human. In the same way, computer generated art has always been even more primal and undeveloped. No service has been able to create a new, wholly original picture that drew from a direct pool of other sources without it being full of artefacts or producing unrecognisable shapes – like dripping oil into a glass of water.
October to December of last year was a perfectly timed period of three months in which a multitude of advanced, open-source, free-to-use AI services became available to the general public. ChatGPT, created by start-up OpenA.I, is an AI algorythm with over 175 billion trained variables, making it the largest and most complex linguistic chat bot to date. If you ask about the origins of the French revolution, it will give you the date, the historical context, as well as the academical discourse and debate related to that subject, all summarised in neatly flowing, conversational language. It’s capabilities are incredibly diverse and well equipped: It can craft poetry, tell children’s stories, produce original lines of code, as well as curate and content control text and passages that you have written.
Similarly, algorithms like StableDiffussion and MidJourney, using a database of millions of photos and art pieces, can generate anything by synthesising pre-existing visual data. It essentially finds the essence of any given style, picture, subject and so on and can extract it and mould it into the given context. An example can be something as ludicrous as “the Cypriot president dressed as Super Mario”, but it can also be used to generate content such as logos or website landing concepts.
AI generated content has been an incredibly divisive topic and has been plagued by misinformation and mysticism. The truth is that, just like the invention of the paper press and the industry of textile machinery, AI is a tool that has slowly been implemented over the years to automate the most banal of writing tasks. Copywriting, ad copy, general press releases and product descriptions have been AI generated for years due to cost effectiveness and time-effectiveness. With the algorithms having the ability to be tailored to specific data pools through machine learning, a firm would be able to incorporate their own language, mannerisms and core values into strenuous writing tasks in a manner of seconds, instead of it taking multiple hours and staff. Practically, a ‘prompt master’ (the person that shapes the requests to generate the right response) would be necessary to act as a moderator, streamlining content creation and work-flow.
AI content can also serve as proof of concept. Specifically with AI generated art, one is able to generate intricate logos for their firm that encapsulate the ethos and message of their branding. This would be done through prompting – example, logo design, flat, vector, high quality, muted colours, business profession – and through diffusion, you have synthesised a new, professional logo for yourself. Similarly, inspiration for artists and graphic designers of any field can be hard to come by due to the complexity of the imagination; AI art is able to generate practically limitless combinations, so reference work can be crafted in a manner of minutes.
Experts have voiced concerns over the repercussions of using these technological hive-minds. In tandem with their abilities, they also have the potential of decimating job markets. With MidJourney being 12 dollars a month, a website developer can simply generate all the stock photos they need, rendering stock photographers redundant. Similarly, graphic designers and portrait artists who labour over their work for weeks now find that they must compete with a quick an easy result that produces staggeringly well-crafted results. Further to this, there is no opt-out system for artists who don’t want their work to be used for AI training. This inevitably will spawn a host of legal issues pertaining to digital ownership and copywrite, with the first lawsuit already being launched by Gettyimages against StableDiffusion for “copying and processing millions of images protected by copyright and the associated metadata owned or represented by Getty Images.”
Furthermore, the sphere of academia has also criticised AI generated content. Chatbots like ChatGPT can synthesise academic essays partially or fully, as well as complete fault proof coding for nearly any task. This has sounded an alarm as it reduces the workload a student might have to do by magnitudes, as well as being used for cheating. ChatGPT can write an entire essay, and because of its unique ability to synthesise information, it is nearly undetectable by plagiarism mechanisms. Philosophy professor Darren Hick from the university of Furman, in South Carolina warned, that “What’s going to be the difficulty is that, unlike convincing a friend to write your essay because they took the class before or paying somebody online to write the essay for you, this is free and instantaneous.”
The paraphysical aspect of AI is a true paradigm shift in the way we produce and consume content, and according to experts, it might eradicate up to 40 per cent of white-collar jobs by the end of the decade, however those would likely be replaced by other tech jobs such as server-ops or development. In reality, when people buy writing content, they look for originality and impact. We must remember that its function is not to create but to synthesise. Originality is still firmly within the grasps of humanity, and as such AI, despite its uncanny-valley level of detail, still does not possess the key elements that differentiates us: Consciousness and originality.
Christodoulos Pouroutidis is an undergraduate in Politics, Philosophy and Law at the University of Warwick