Artificial Intelligence (AI) is changing the way we do everyday tasks and activities. In this blog, we answer some common questions about AI to help you understand the basics.
Machine Learning is a technique of AI that focuses on developing algorithms (sets of rules or instructions) that enable systems to learn from data, identify patterns, and make decisions outside of human control. Example: Suggesting films to watch next on Netflix based on viewing history.
Deep Learning is a specialised technique within Machine Learning that uses neural networks (designed to mimic the human brain) to analyse large amounts of data and carry out tasks commonly done by humans. Example: Recognising faces in photos or translating languages.
NLP is a branch of AI that allows machines to communicate in real time in any human language, either verbally or in writing, to respond to tasks they are designed to complete. Examples: Chatbots like ChatGPT or assistants like Copilot, which can generate human-like responses in conversations. NLP can also describe images or answer questions about them or create pictures or video content.
Predictive Analytics involves applying statistical methods, such as regression analysis (estimating trends over time), and algorithms, such as decision trees (making choices based on conditions), to anticipate future outcomes based on historical data. Example: Predicting weather patterns or identifying when industrial machinery, such as factory equipment or aircraft engines, might require checking and changing of parts.
Big Data refers to the vast amounts of data, often measured in terabytes or petabytes, that AI uses to operate effectively. Structured data includes organised formats like spreadsheets or databases, while unstructured data might be videos or social media posts. AI analyses this data to find patterns and make recommendations or improvements
Bias occurs in AI systems when the data used reflects human preferences, unreliable data, like incomplete records, or flawed algorithms, like coding mistakes. This affects fairness and accuracy. Example: Hiring algorithms used by recruitment companies wrongly prioritising certain people based on age or ethnicity.
Explainability refers to understanding and interpreting how AI models make decisions. Example: In loan approvals, it could involve showing which factors, such as income or credit score, influenced the decision to approve or reject a loan application.
Interoperability is the ability of AI systems and devices to share data and work well together. Example: Connecting a fitness tracker to a smartphone app that tracks your activity and suggests simple goals for improving your health.
