What actually is Artificial Intelligence
When you hear the term ‘Artificial Intelligence’, what comes to mind? Chatbots? LLMs? Obnoxious posters? Or perhaps some commonly used tool names? However, Artificial Intelligence did not begin with ChatGPT. Matter of fact, the very concept of creating something able to imitate human intelligence is far older than any computer.
Humans have been fascinated with mechanical structures for millenia. Philo of Byzantium designed the ‘automatic maid’; a machine supposedly able to pour water and wine in a cup it would hold, back in 250 BC. Subsequently, for ages, they have also imagined mechanical beings, automata, and systems capable of reasoning. By the 20th century, these hypotheses shifted from philosophical debates to mathematical and computational problems. Alan Turing, in 1950, famously raised the question whether machines could ever be capable of thought. A few years later, in 1956, at Dartmouth College, the term ‘Artificial Intelligence’ was coined, and became an emerging academic field.
Thus, what exactly is AI?
In simple terms, AI is an umbrella term encompassing computer systems designed to, and most importantly, capable of, performing tasks we usually associate with ‘intelligent’ behaviour: recognising patterns, solving problems, or even generating content. And this is where a usual confusion occurs.
AI and Machine Learning are NOT the same thing.
AI is the broader field. Machine Learning is one AI approach where computers learn patterns from data; rather than adhering to a set of explicit preprogrammed rules. Moreover, this brings us to deep learning. The latter is a subset of Machine learning, where multi-layered neural networks identify even more complex patterns. We moreover have generative AI, which is the talk of the town nowadays. This type of AI, including tools like ChatGPT, creates content; for example text, images or even videos.
In a nutshell, AI has been around way longer than the recent generative AI boom, and people have been using it everyday without necessarily calling it AI. Notwithstanding, no matter how impressive these tools get with every passing day (or even moment, in this fast-paced technology sector), this does not necessarily make them infallible. AI systems can inherit biases based on their training data, provide wrong output with full confidence (remember when ChatGPT would confidently say there are only 2 “r’s” in the word “Strawberry”?), and provide results that might be correct, yet difficult to assimilate and actually discuss.
As these systems become increasingly popular in a myriad of fields (from IT and finance to healthcare), it is now, more than ever, important to ensure that everyone properly understands what Artificial Intelligence actually is. And yet, more than 70 years after AI became a formal field of study, that question is somehow still surprisingly hard to answer. Perhaps, understanding that uncertainty, is far more important than understanding the technology itself. After all, Socrates once said “what I do not know, I do not think I know either”, and this is exactly the mentality we should adopt when it comes to new technologies.
