AI: What is it really?
Artificial intelligence (AI) is an umbrella term for data-analysing systems that can generate forecasts, recommendations and decisions that are all based on probabilities. Its operations are not restricted by any set mechanical rules, and so AI is able to ‘learn’ and apply its learning. Remember, however, that AI does err at times.
Everyday examples of AI use: In spam filters, journey planning software, or systems that create sales forecasts based on previous sales-related data.
Generative artificial intelligence (GenAI) is a type of AI that specialises in creating new content (text, images, computer code, etc). It generates responses based on probability calculations. GenAI always aims to furnish a slick answer, but therein lies the danger of ‘hallucinations’, or its tendency to simply makes up content in order to give an answer and sound convincing. Therefore, the results obtained from AI must always be critically evaluated.
Everyday examples: Using Microsoft Copilot or ChatGPT to summarise long documents, draft emails or translate languages.
AI agents to a certain degree, can act autonomously. While GenAI only gives answers to the questions that are asked, AI agents are able to perform more goal-oriented behaviour, breaking down a large assignment into stages and employing various tools (like browsers or other software) to get the individual tasks done.
Everyday examples: An AI agent may be given the following assignment: ‘Find three different grocery stores and compare the price per kilo of apples. Make a table showing the price comparisons at the different stores and send a summary of it to me by email.' The AI agent performs the various steps needed for the assignment without further human assistance.