Artificial Intelligence

Automated Decision Making

• Data Analysis: The system ingests structured or unstructured data to build a foundational understanding of the problem domain.
• Rule Application: Predefined constraints and learned heuristics are applied to the data to filter possibilities.
• Optimization: The system evaluates potential actions against a cost function or goal (e.g., minimizing delivery time or maximizing accuracy) to select the best path.

Machine Learning (ML)

• ML is a subset of AI that focuses on the use of data and algorithms to imitate the way that humans learn, gradually improving its accuracy. It relies on training data to build models that can make predictions on unseen data.

• Automation vs. AI: While basic automation follows a fixed script to perform repetitive tasks, AI uses reasoning to handle variability and make subjective-style judgments.

• Identify the 'Why': When asked about AI applications, always link the technology to a specific human cognitive function it is replacing or enhancing (e.g., 'AI in medical imaging replaces human visual perception to detect anomalies').

• Evaluate Ethical Trade-offs: Be prepared to discuss the balance between efficiency (speed of AI) and accountability (who is responsible when the AI makes a mistake).

• Check for Bias: Always consider the quality of the training data. If the input data is biased, the AI's output will inevitably reflect and amplify that bias, regardless of how sophisticated the algorithm is.

• Human-in-the-loop: Look for scenarios where human oversight is necessary, especially in high-stakes environments like healthcare or legal sentencing.

• The 'Objectivity' Myth: A common misconception is that AI is perfectly objective because it is a machine. In reality, AI inherits the biases present in its training data or the design choices of its creators.

• General vs. Narrow AI: Students often confuse 'Narrow AI' (systems designed for one specific task, like chess or facial recognition) with 'General AI' (hypothetical systems with human-like consciousness). Currently, all functional AI is Narrow AI.

• Correlation vs. Causation: AI is excellent at finding correlations (things that happen together), but it does not inherently understand causation (why one thing causes another).