What is Artificial Intelligence?

Artificial Intelligence is poised to shake-up the market in a big way, paving the way for incredible leaps in our ability to process and understand data, create content and automate some of the more tedious everyday tasks. But what do people mean when they talk about AI? Can a computer really think like a human being?

Fortunately, we’re a long way away from sentient computers. Though tech entrepreneurs are reaching in that direction in the long term, the artificial intelligence in use today is known as artificial narrow intelligence, characterised by its ability to complete a specific task for which it’s programmed. Stronger forms of AI, such as that which is on par with or superior to human intelligence, is still in the realm of theoretical.

But don’t let the term ‘narrow’ fool you — AI is capable of some pretty sophisticated feats. Voice assistants like Siri and Alexa can interpret spoken words into functional commands, quickly responding to a wide variety of queries. Chatbots work similarly with the written word, using an algorithm to provide responses in a conversational tone. Even the technology powering autonomous vehicles takes advantage of AI, using pattern-matching algorithms to identify road hazards and traffic signage.

At its core, AI relies on identifying and categorizing data; one way this is accomplished is through machine learning. Machine learning uses mathematical models of data to help a computer ‘learn’, letting the computer use its own results to improve upon its output. As the computer processes large quantities of data, its results can be optimised over time. A neural network is a type of machine learning model that is mimics the way biological neurons work in the human brain. The model can identify various phenomena, assign weights to different options and arrive at a conclusion based on all of these factors, much the same way a human would.

While there is great promise in the practical application of AI technologies, the deployment hasn’t been without its share of speedbumps. AI requires large amounts of data to appropriately train its models, and numerous questions have been raised about the types of data these models are harvesting, including legal and ethical issues. Although chatbots such as OpenAI’s ChatGPT can respond in well-crafted sentences, there have been issues with ‘hallucination’ — citing facts that don’t exist. While the potential is there, there are still some bugs and irregularities to be worked out in these systems.  

Ultimately, AI should be regarded as a powerful and useful tool, not as a threat that should be avoided. And as these tools become more sophisticated and common in business settings, it’s important to evaluate your own unique business needs to see where you can take advantage of this fascinating new technology.