What is machine learning (ML)?

A type of AI where systems learn patterns from large amounts of data rather than being explicitly programmed with rules — improving their performance through experience Software that can learn new programming languages and write code autonomously Computers that learn faster the more RAM they have A system that memorizes all available information and retrieves it on request

What is a 'neural network' in AI?

A computational model loosely inspired by the human brain — consisting of layers of interconnected nodes that process data and learn to recognize complex patterns A network of connected AI systems that communicate to solve problems collectively The internet infrastructure that connects AI data centers around the world Brain-computer interface hardware that connects human neurons to computers

What is a 'large language model' (LLM) like the ones powering AI chatbots?

An AI trained on vast amounts of text to predict likely next words/tokens — enabling human-like text generation, conversation, summarization, and reasoning A database containing all information from the internet that AI can look up A programming language developed specifically for building AI systems A computer program that has memorized dictionaries in multiple languages

What is 'AI hallucination' and why does it matter?

When an AI generates confident-sounding but factually incorrect information — it matters because users may trust and act on wrong information without realizing it's false When AI systems start behaving randomly due to errors in their training data Visual glitches that AI image generators produce in generated images When AI chatbots go offline unexpectedly due to server problems

How is AI being used in cybersecurity — both offensively and defensively?

Defensively: AI detects anomalies and threats faster than humans; offensively: AI helps attackers craft better phishing, find vulnerabilities, and automate attacks at scale AI is only used defensively — attackers don't have access to advanced AI tools AI is only useful for attackers — defenders still rely on traditional rule-based security AI has had no significant impact on cybersecurity — humans are still more effective

What is 'Internet of Things' (IoT) and what security risks does it create?

IoT is everyday objects connected to the internet (smart TVs, baby monitors, thermostats, cars) — they create security risks because they often have weak security and hard-to-update software IoT only refers to industrial sensors in factories — not consumer devices IoT devices are more secure than computers because they have simpler software IoT doesn't create security risks because the devices don't store personal data

What is 'cloud computing' and what are its security implications?

Using remote servers over the internet for storage, processing, and software — the provider is responsible for physical security, but data security and access control remain shared responsibilities Cloud computing is inherently less secure than on-premise servers because data leaves your building The cloud provider is fully responsible for all security — customers don't need to worry about it Cloud computing only refers to online file storage like Dropbox or Google Drive

What is a 'smart city' and what privacy concerns does it raise?

Cities using networked sensors, cameras, and data analytics to improve services (traffic, energy, safety) — raising concerns about mass surveillance, data ownership, and who controls the infrastructure Cities where all residents have free internet access provided by the government A marketing term for cities that have attracted large technology company headquarters Cities that require residents to use smartphone apps for government services

What is 'autonomous vehicle' technology and what cybersecurity challenges does it introduce?

Self-driving cars that use sensors, cameras, and AI to navigate without human input — cybersecurity is critical because a hacked vehicle could be weaponized, steered into traffic, or have its brakes disabled Autonomous vehicles have no cybersecurity risks because they don't connect to the internet Only the navigation systems need cybersecurity — brakes and steering can't be hacked Autonomous vehicles are more secure than traditional cars because they remove human error

What is 'biometric authentication' and what are its privacy trade-offs?

Using unique physical characteristics (fingerprint, face, iris, voice) to verify identity — convenient and hard to fake, but biometric data is permanent and a breach is irreversible unlike a password Biometrics are the most secure form of authentication and have no significant downsides Biometric authentication is less secure than passwords because it can be spoofed with photos Biometrics are only suitable for physical access control, not digital authentication