Ethics, IT Act 2000 and AI notes

Computer ethics is the set of principles that say what is right and wrong when you use a computer, a network or data. The machine is fast and has no feelings (Unit 1). That is why a person must bring the rule. Copying a classmate’s assignment file, opening someone else’s mailbox, or running a lab PC as if it were a toy that cannot harm are ethics questions before they are law questions. Law (next lesson) can punish; ethics is the habit that should stop the act even when no invigilator is in the lab.

The books gather a short list. Do not use a computer to harm other people or their work. Do not snoop in files, mail or mark-lists that are not yours. Do not steal time, bandwidth, software or data (piracy is the named form in the next lessons). Do not bear false witness with a computer: a forged mark-sheet, a fake notice in Anil’s name, a rumour mailed as fact. Respect property — including programs and writing, which are intellectual property. Think about the social effect of what you publish: a lab photograph, a comment on a classmate, a file dumped on a public folder.

Privacy, accuracy, property and access (sometimes called PAPA) are four doors the same books use. Privacy: who may see roll 2145’s marks. Accuracy: a wrong entry on a result server is not “the computer’s opinion.” Property: who owns the notes file. Access: who is allowed to log in. Ethics is not a fifth generation of hardware and not a virus scanner. It is the human rule for the tools of Units 1–4.

Cyber law is the law that covers acts done with computers, computer systems, computer networks and data in electronic form. Ordinary law still applies (theft, defamation), but bits on a wire needed a statute that said an electronic record can be as valid as paper, and that named computer offences. In India that statute is the Information Technology Act, 2000 (IT Act 2000), in force from 17 October 2000. An amendment in 2008 added and tightened several offences. This is an introduction, not the full Act.

The Act gives legal recognition to electronic records and to digital signatures (later also electronic signatures): a filled exam form or a notice on an official site can stand in law the way a signed paper can, if the conditions of the Act are met. It names a Controller of Certifying Authorities for digital-signature certificates. It defines computer, computer network, data, electronic record — the vocabulary Unit 1 already used in plainer words.

Offences the introductory books attach to the Act include unauthorised access and damage to a computer or network, spreading a virus or other contaminant, dishonestly hiding or changing source code, publishing certain kinds of obscene electronic material, and identity-related cheating. Punishment is fine and/or imprisonment as the sections state. The lab PC is not outside the law.

Cyber law is not the same as computer ethics. Ethics is the ought; the Act is the must, with a court. E-governance and e-payment from Unit 4 sit on this legal floor: without recognition of electronic records, an online result or an EFT message would not count.

This unit names four wrongs the IT Act (and related law) touch: four ways a computer can injure property or a person.

Software piracy is the unauthorised copying, use, distribution or sale of a program. Installing one licensed Rapidex on every lab PC without the licence that allows it, sharing a paid Windows installer on a pen drive, or selling burned CDs of an office suite are piracy. Softlifting (using a single-user copy on many machines), counterfeiting (fake boxes), and internet piracy (upload/download of paid software) are the forms books list. A pirated copy can also carry a virus. Ethics said do not steal programs; the Act and copyright law give that theft a court.

Intellectual property (IP) is the legal name for creations of the mind: a program, a book, a mark-list design, a college logo. Copyright protects original writing and software from unauthorised copy. A patent can protect a new invention (including some technical methods). A trademark protects a name or sign that identifies a product (a software brand). Copying a classmate’s HTML page and putting your roll on it is a copyright wrong as well as an ethics wrong. IP is why Unit 3’s “simple HTML program” you write is yours, and why you may not zip the whole of Sinha for the class.

Child abuse in this unit is the IT Act’s concern with using a computer or network to store, publish, transmit or lure in connection with sexual content involving a child, or to target a child for harm online. The 2008 amendment sharpened these offences. The notes name the crime and the duty: do not create, keep, share or seek such files; report through the college or the police if such material appears on a lab PC. This lesson does not describe the content.

Social engineering is tricking a person into giving access, a password, a roll number, a card detail or a click — not breaking the CPU. The computer is honest; the human is persuaded. Phishing is a fake e-mail or page that looks like a bank, a university result portal or a scholarship site (“confirm roll 2145 and your password”). Pretexting is a phone or message in a false role (“lab technician, send your login”). Baiting is a “lost” pen drive in the lab that installs harm when opened. Defence is the habit from lesson 1: do not give marks, passwords or OTPs to a page or a person you did not seek; check the URL; ask the college office, not the mail.

What is AI? Artificial Intelligence is the branch of computer science that tries to make machines do tasks that, when a person does them, we call intelligent: reason, learn from examples, understand language, see, plan. John McCarthy’s classroom wording (1956) is programs that behave as if they have intelligence. It is not “any fast computer.” A lab PC adding 36 and 64 is arithmetic, not AI. A program that suggests the next word, finds a face in a photograph, or chooses a route around a jam is closer to the AI the books mean. Unit 1’s fifth generation is hardware and an aim; this lesson is the field itself.

History. Alan Turing (1950) asked whether a machine could imitate a person’s answers well enough to pass a test. The name AI is tied to the Dartmouth meeting, 1956, and McCarthy. Early programs played checkers or proved small theorems. The 1980s brought expert systems (rules from a human expert, used in medical or technical advice) and Japan’s fifth-generation project (Unit 1). Later work added learning from data, speech and vision. The names are Turing, the 1956 meeting, expert systems, and the fifth-generation link.

Applications. Expert systems; robotics (a machine that senses and acts); natural language (machine translation, simple question-answering); computer vision (reading an OMR sheet is optical mark recognition from Unit 1 — related, not the whole of vision); medical diagnosis support; games; teaching drills that adapt to a student’s wrong answer. A bank that flags a strange payment, a map that reroutes, a spam filter on e-mail — applications on the same list.

Advantages. Speed on huge search spaces (weather models already sat on supercomputers in Unit 1; AI methods help the search). The same advice at 2 a.m. as at noon (diligence). Can watch more cameras or more lab logs than one clerk. Can open a tool to people who cannot travel to a specialist. Disadvantages. No common sense unless we have encoded or taught it; garbage in, garbage out still applies. Cost of data and machines. Errors can look confident. Jobs change; some routine work shrinks. Privacy: a vision system that reads faces in a campus is also a watch. A program does not carry ethics by itself (lesson 1 of this unit).

AI in day-to-day life. Predictive text and auto-correct on a phone; voice assistants that set an alarm; maps that pick a road; spam filters; a shopping site that shows another book after Rapidex; face unlock on some phones; captions on a video. Checking an MGSU notice in a browser is not itself AI. Asking a voice tool to read that notice aloud is the daily face of the field. The fifth generation’s “natural-language aim” is this: talk to the machine with less rigid programming. The limits of Unit 1 still hold: the machine has no feelings and no unprogrammed rule.