Artificial intelligence (AI) is something of a buzzword presently, the world has been electrified with the possibilities and downfalls that AI can bring humanity. AI as a topic of conversation is hardly new, the concept of autonomous vehicles (reliant on AI) has been on the horizon for years, however, the launch of Chat GPT has caused the world to spiral into a full-blown hyper fixation. When looking up on Google (as of 1/04/24) “AI news”, the results provide further question suggestions, such as “Will AI overtake jobs?”, “What jobs AI can’t replace?” and “Which jobs are most at risk from AI?”. Characterising AI as a malicious threat. This caricature of AI is further exacerbated by media such as “The Matrix” (1999) in which computer systems fully control humanity and “The Incredibles (2004)” in which a “learning robot” breaks free of the villain’s control to destroy a city. Society has overestimated AI and perceives it to be a menace.
According to the Oxford Dictionary, AI is defined as “The theory and development of computer systems able to perform tasks normally requiring human intelligence, such as visual perception, speech recognition, decision-making and translation between languages.” This is another way of saying AI is a computer system that can perform a task complex enough that we associate it with humanity. Whilst a computer system may be able to perform a singular task, current technology is not advanced enough that we can fully equate it to being human.
A simple predictive model using linear regression is by technicality a form of AI, however, it is merely calculating a line of best fit to a set of data and utilising that to inform future decisions. Another example is a decision tree, a code can make informed decisions depending on inputted factors and comparative conditions, for example, a decision tree could be used to decide whether a user can drive, if the user’s age is greater or equal to 17 then, if the user can read a number plate from 20 feet and so on. AI merely is complex mathematics.
Whilst AI is far simpler than society perceives it to be, the issues created by it are far more complex beyond whether it will take a nation’s jobs. Whilst AI does pose a threat to the creative arts industry it will also result in an echo chamber (AI art is created and published, AI is then trained on such art and henceforth only produces such art) which only human intervention can break. One of the more pressing issues surrounding AI is that it reinforces societal biases. AI is only as clever as the data it has been trained upon. As of 04/04/2024, when asked an image-generating AI for images on the prompt “Computer scientist”, it produced images of white men in lab coats and glass, perpetuating gender norms and erasing women and people of colour’s important contributions to the field of computer science. Joy Buolamwini, author of Unmasking AI and founder of the Algorithmic Justice League, provided further evidence of the stereotypes evident in AI algorithms, noting that facial recognition software has a lower accuracy rating when identifying people of colour, more specifically women of colour. She goes on to further discuss the real-world impact of these consequences in her paper (co-authored by Timnit Gebru) “Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification” in which she highlights how people could be wrongly convicted of crimes due to erroneous AI identification of camera footage. AI’s reinforcements of stereotypes have far greater consequences than the possibility of AI stealing people’s jobs.
As with most branches of AI, Chat GPT has also been subject to misinterpretation. Chat GPT is a chatbot and a large language model, invented by OpenAI and launched on 30/11/2022. Chat GPT has since been an enormous success, and companies such as Bing, Google and Grammarly have strived to create similar functions. Being a large language model Chat GPT statistically analyses the probability of a given word following another provided with a prompt. It cannot think and has also been overused as a source of factual evidence and mathematical calculations. For example, when asked for a quote from Macbeth by William Shakespeare on the theme of light and darkness (01/03/2024), it presented “Light, seeking light, doth light of light beguile; So, ere you find where light in darkness lies, Your light grows dark by losing of your eyes.” claiming it is spoken by Malcolm, King Duncan’s oldest son in Act 1 Scene 4, but is instead a quote from Act 1, scene 1 of Love’s Labour’s Lost also by Shakespeare. It is unable to fact-check the information it is stating and merely provides a probability analysis of the English language. Despite this, it is still heralded as a polymath.
The release of Chat GPT has furthermore launched an AI craze, however, AI has been in our lives for much longer than commonly thought. AI conceptually began with Alan Turing in his unpublished paper “Intelligent Machinery” in 1948 and in 1951 the first AI computer program was programmed and run. AI exists in day-to-day life from the algorithms that produce the order of webpages, and the autofill on phone texting. All translation software such as Google Translate and DeepL use AI, Google Translate being older than the iPhone and DeepL being released over 6 years ago. This article itself has been written with the assistance of AI, Microsoft word’s built-in Editor uses AI as well as search engines required to perform the research.
In conclusion, the characterisation of AI portrayed in media is a caricature, often stemming from an underestimation of what classifies as AI and a misunderstanding of the key issues of the issues surrounding AI, which, by extension allows the perpetuation of societal issues such as racism and misogyny. Chat GPT is being overestimated in its abilities and society is underestimating the age and history of AI, perceiving it to be a relativity modern technology, when it is over 75 years old.
