1. Learning outcomesExplanation
2. First: computer networks and the Internet / intranet infrastructureExplanation
3. The local area network (LAN)Explanation
4. The wide area network (WAN)Explanation
5. Second: cybersecurity, digital privacy and fighting threatsExplanation
6. Phishing and smishingExplanation
7. RansomwareExplanation
8. Social engineeringExplanation
9. Third: the smart educational assistantExplanation
10. Prompt engineeringExplanation
11. Verification, credibility and anti-impersonationExplanation
12. Worked model itemsExplanation- Describe the network infrastructure behind the Internet and the intranet, and compare a LAN with a WAN.
- Explain cyber threats — phishing, smishing, ransomware and social engineering — with their detection and prevention.
- Use a generative-AI educational assistant for summarising, explaining and coding, and write an accurate prompt.
- Verify AI output, recognise hallucination, and keep academic integrity and intellectual-property rules.
🌳 Concept tree
Networks, Cybersecurity and Artificial Intelligence ├── First: computer networks and the Internet / Intranet infrastructure │ ├── LAN characteristics │ └── WAN characteristics ├── Second: cybersecurity and the fight against threats │ ├── phishing and smishing │ ├── ransomware │ └── social engineering └── Third: generative AI, cloud computing and technology ethics ├── the smart educational assistant ├── prompt engineering └── verification, hallucination and plagiarism
01 Learning outcomes
- Relate a networking need to the right class of network and justify the choice.
- Name the most common attacks, the psychological trick each one uses, and its technical prevention.
- List the tasks an AI assistant can perform in studying, and the tasks it must not be trusted with blindly.
- Compose a prompt that fixes role, context, task, constraints and output format.
02 Key terms
- LAN — Local Area Network over a small area. WAN — Wide Area Network over large distances.
- Phishing / smishing — credential theft by fake e-mail or SMS.
- Ransomware — malware that encrypts files for a payment.
- Social engineering — psychological manipulation of people rather than of machines.
- Generative AI — models that produce text, images or code in response to a prompt.
- Prompt — the instruction given to the model; prompt engineering is writing it precisely.
- Hallucination — the model stating something false, or inventing references, with full confidence.
03 First: computer networks and the Internet / intranet infrastructure
- Computer networks are the physical and logical infrastructure over which Internet services travel: devices, links, routers and protocols.
- The Internet is the public open network that connects billions of users and whose content anyone may reach.
- The intranet is a private closed network limited to the members of one institution — bank staff or school students — and cannot be entered from outside except through encrypted security clearance.
- Both are built with the same technologies; the difference is the access policy, not the cables.
04 The local area network (LAN)
- Covers a small, limited area: one building or one campus.
- Characteristics to quote in the exam: high speed, very high data-transfer rate, easy management, and strong security because the media are under one administration.
- Typical use: sharing printers, local files and school servers inside one site.
05 The wide area network (WAN)
- Covers vast areas — cities and continents — using communication lines, satellites and submarine fibre-optic cables.
- Costlier and harder to manage than a LAN, and generally lower in local transfer speed because of the long links and the intermediate equipment.
- The Internet is the largest WAN; a bank linking its branches abroad runs a private WAN.
- Comparison table for revision: coverage (small / vast), speed (very high / depends on the carrier), management (one administration / several providers), cost (low / high), example (school lab / the Internet).
06 Second: cybersecurity, digital privacy and fighting threats
- Cybersecurity is the practice of protecting systems, networks and data from attack; digital privacy is the user's right to control his own data.
- Threats combine a technical vector (a link, an attachment, a vulnerable program) with a human weakness (trust, curiosity, fear) — which is why defence is partly behaviour and not only software.
- The standard protective set: strong unique passwords, multi-factor authentication, timely updates, original antivirus with an enabled firewall, downloads from official stores only, and periodic independent backups.
07 Phishing and smishing
- Phishing: fake e-mails that appear to come from a bank or a famous company and contain malicious links urging the user to log in and steal his data.
- Smishing: the same scheme sent by SMS.
- How to detect it: inspect the sender's e-mail address carefully, note spelling mistakes, distrust urgent demands for secret data, and hover over links to see the real destination.
- Correct action: do not click, do not reply; open the site by typing its address, then report the message.
08 Ransomware
- Harmful programs that encrypt the computer's files completely; the attacker then demands a money payment for the decryption key.
- Entry points: an attachment opened from an unknown sender, a fake installer, or an un-patched vulnerability.
- Prevention as laid down in the syllabus: keep software updated, never open unknown attachments, and take periodic backups that are stored independently (offline or in the cloud with versioning), so the machine can be restored without paying.
09 Social engineering
- The psychological manipulation of people, exploiting trust, curiosity or fear to make them disclose secret information — the human being, not the machine, is the attacked system.
- Forms: a caller pretending to be technical support asking for an OTP, a prize message that requires "verification", a fake recruiter asking for ID documents, tailgating into a restricted room.
- Defence: verify the identity of the requester through an official channel, refuse to reveal codes, and apply the golden rule — no bank or official institution ever asks for your password or CVV/OTP.
10 Third: the smart educational assistant
- Generative AI supports learning by: summarising long texts, explaining and simplifying complex concepts, suggesting research ideas and outlines, and writing code and correcting its errors.
- Good practice: use it as a tutor that explains and questions you, not as an answer machine that writes your assignment.
- It works with cloud computing, so the heavy processing runs on remote servers rather than on the student's device.
11 Prompt engineering
- The quality of the AI output depends completely on the accuracy and clarity of the prompt written for it.
- A complete prompt states: the role ("act as a physics teacher"), the context (the grade and the topic), the task (what to produce), the constraints (length, level, language) and the required format (table, bullets, numbered steps).
- Iterate: refine the prompt with feedback instead of repeating the same vague request.
- Vague prompt → generic answer; this cause-and-effect sentence is a model answer for "explain why the output was poor".
12 Verification, credibility and anti-impersonation
- AI answers must never be taken as established facts without checking, because models suffer from hallucination — inventing false information and even non-existent references.
- Cross-check every fact against a trusted source: a textbook, a published paper or an official site; verify each cited reference actually exists.
- Respect intellectual-property rights and avoid academic plagiarism; disclose the use of AI tools instead of presenting generated work as your own.
- Beware deepfakes and impersonation: confirm an identity claim through a second channel before acting.
13 Worked model items
- "A school lab needs a fast, easily managed and secure network inside one building." → a LAN, justified by small coverage, high transfer rate, easy administration and security.
- "An employee received an SMS saying his card is blocked and asking for the OTP." → smishing plus social engineering exploiting fear; the answer: never share the OTP, contact the bank through its official number.
- "After opening an attachment, all files became unreadable with a payment note." → ransomware; the recovery path is restoring the independent backup, not paying.
- "Write a prompt to get a revision table of the layers of the TCP/IP model." → role + context + task + constraints + format, e.g. "Act as an ICT teacher; for grade 1 make a four-row table with layer, its job and one protocol per layer, in 120 words."
14 Exam traps and key takeaways
💡 High-yield traps
- LAN versus WAN is a question of coverage; speed and manageability follow from it — quote the syllabus wording.
- An intranet is not a "smaller Internet" technically; it is the same technology with closed access.
- Social engineering attacks the person; malware attacks the machine. Questions often ask which one is described.
- Ransomware demands a payment for decryption; spyware steals credentials silently — different goals, different defences.
- Backups must be independent (offline/separate) or the ransomware encrypts them too.
- Prompt quality controls output quality — "the model is wrong" is not the expected answer for a poor result.
- Hallucination includes invented references, not only wrong numbers.
- Using AI is allowed; submitting generated work as your own without disclosure is plagiarism.