Chapter Practice + Test Paper (AI Ethics ka Final Round!)
Poore AI Chapter 3 ki ek-page revision, confuse karne wali jodiyan, ethics-dilemma solve karne ka tareeka, Python mini project, aur 40 marks ka model test paper: sab answers ke saath 📝🤖
🧭 Is Post Mein
📖 Kahani: “Ethics toh bas lambi baatein hain!”
Library mein Rohan ne kitaab band karte hue kaha, “AI ka chapter toh coding wala hona chahiye tha. Ethics toh bas lambi-lambi baatein hain, test mein kuch bhi likh denge!” 😅
Golu pareshan tha: “Mujhe toh morals aur ethics mein hi confusion hai. Phir 6 categories, data ke 6 niyam, policy ke 8 hisse, 7 principles… sab ek jaisa lagta hai!”
Priya ne notebook kholi: “Dekho, poora chapter ek hi sawaal ka jawaab hai: AI se kisi ka nuksaan na ho, aur faayda sabko mile. Iske liye pehle samjho ki sahi-galat kaun tay karta hai (morals, ethics), phir kya galat ho sakta hai (categories, data, access), phir use kaise rokein (policy, principles). Aur ethics ke sawaalon mein ‘kuch bhi’ nahi chalta: kaaran chahiye, sahi shabd chahiye.”
Teeno ne ek page ka chapter map banaya, phir ghadi lagakar model test paper solve kiya. Aaj hum bhi wahi karenge. Taiyaar? ⏱️
🎯 Is Post Mein Kya Milega (aur kaise use karein)
- Chapter Map: AI Chapter 3 ke saare topics ek table mein: matlab, yaad rakhne ki trick, exam ka keyword
- 3 diagrams: chapter ka mind map, “ethics dilemma kaise solve karein” flowchart, aur real examples ki timeline
- Sabse zyada confuse hone wali jodiyan (Morals vs Ethics vs Law, Personal vs Sensitive data, Principle vs Policy…) aur galatfehmiyan
- Python Mini Project: “AI Ethics Category Finder”, jo situation padh kar uski category batata hai (AI Tutor mein chalega)
- Warm-up (Fill in the Blanks, True/False) → Model Test Paper (40 marks, 1 ghanta) → Output sawaal → HOTS → Cyber Olympiad
Kaise use karein: Pehle A se D tak padho (25–30 min). Phir copy-pen lekar Test Paper khud solve karo, ghadi lagakar. Answer tabhi kholna jab apna jawaab likh lo. Aakhir mein marks gino aur kamzor topic ka post dobara padho. 💪
🧱 Basics: AI Chapter 3 ke Saare Posts (revision links)
Agar test mein koi topic kamzor lage, toh yeh post dobara padho:
- AI Ethics Kya Hai?: ethical scenarios, AI ethics ka matlab, kyun zaroori, zimmedari kiski, 5 Sawaal Ethics Check
- Morals vs Ethics aur AI Ethics ke Prakar: morals, ethics, law; AI ethics ki 6 categories
- AI Access aur Personal Data: digital divide, access ki chintaayein, personal/sensitive data, data ke 6 niyam, DPDP Act 2023
- Achhi AI Policy aur Real Examples: policy ke 8 hisse, galtiyon ke kisse, achhe examples, risk-based approach
- AI Ethics ke Principles aur Future: NITI Aayog ke 7 principles, OECD, UNESCO, 5 classic siddhant, future aur why it matters
💡 Concept: Poora Chapter Ek Table Mein (Chapter Map)
Ethics matlab sahi-galat batane wale niyam aur soch. AI Ethics matlab AI ko sahi tareeke se banane, train karne aur use karne ke siddhant. Kyun zaroori? Kyunki AI bahut tez aur bahut bade paimaane par faisle karta hai, data se seekhta hai (toh data ka bhedbhaav bhi seekh leta hai), aur use khud sahi-galat ki samajh nahi hoti.
| Topic | Ek line mein | Yaad rakhne ki trick / keyword |
|---|---|---|
| AI Ethics | Technical sawaal “bana sakte hain?”; ethical sawaal “banana chahiye?” | AI ek tool hai; zimmedari insaan ki (Developer, Company/School, Sarkar, User) |
| 5 Sawaal Ethics Check | Kisi bhi AI ko jaanchne ke 5 sawaal | Fayda/Nuksaan · Barabari · Ijazat/Privacy · Saaf-safaai · Zimmedari |
| Morals vs Ethics vs Law | Apne vishwas / group ke saajhe niyam / sarkar ke niyam | Morals = man ki awaaz · Ethics = group ka code · Law = saza wala niyam |
| 6 Categories | AI ki samasyaon ke 6 prakar | Bias & Fairness · Privacy & Data · Transparency · Accountability · Safety & Security · Social Impact |
| AI Access | AI tools tak pahunch: device, internet, bhasha, gyaan, paisa | Digital Divide = jinke paas hai aur jinke paas nahi, unke beech ki khaai |
| Personal Data | Jisse aapki pehchaan ho sake | Sensitive: password, OTP, bank, health, biometric, Aadhaar |
| Data ke 6 niyam | Data ko sahi tareeke se lena aur rakhna | Consent · Purpose Limitation · Data Minimization · Security · Storage Limitation · Rights |
| DPDP Act, 2023 | Bharat ka digital personal data kanoon | Consent zaroori; 18 se kam umr: maata-pita ki verifiable consent |
| AI Policy | AI ke use ke likhe hue niyam | 8 hisse: Uddeshya · Kya karein/na karein · Fairness · Privacy · Transparency · Human Oversight · Zimmedari & Shikayat · Safety, Testing & Review |
| Risk-based approach | Jitna khatra, utne sakht niyam | Unacceptable (mana) → High → Limited (batana zaroori) → Minimal |
| NITI Aayog ke 7 principles | Responsible AI (Feb 2021) | SEI-PTA-H: Safety & Reliability · Equality · Inclusivity · Privacy & Security · Transparency · Accountability · Positive Human Values |
| 5 classic siddhant | Duniya bhar mein maane jaane wale | Beneficence · Non-maleficence · Autonomy · Justice · Explicability |
| Future | Aage ki chunautiyan aur koshishein | Deepfakes · naukriyaan · Green AI · bachche aur AI · summits (New Delhi, Feb 2026) |
🖼️ Diagrams: Picture se Samjho
🔐 Deep Concept: Chapter ki 3 Seedhiyan (Kaun → Kya → Kaise)
Chapter mein bahut saari lists hain, par sab ek seedhi ki tarah judi hain. Yeh samajh liya toh “kyun” wale sawaal aasaan ho jaate hain:
Exam tip: koi bhi case aaye, teeno seedhiyan chadho: kaun zimmedar, kya galat hua (category), kaise theek ho (principle + policy ka hissa). 🎯
✏️ Ethical Scenarios: Situation → Category → Principle → Hal
1. School ka face-attendance camera saanwle rang ke bachchon ko baar-baar “absent” dikhata hai.
Category: Bias & Fairness. Principle: Equality / Justice. Kyun: AI ko zyada tar gore chehron ke data par sikhaya gaya hoga (Gender Shades jaisa). Hal: har tarah ke chehron ka data, testing, aur teacher ka manual check (human oversight).
2. Ek free torch app contacts, location aur microphone ki permission maangti hai.
Category: Privacy & Data. Niyam: Data Minimization aur Purpose Limitation tootey. Kyun: torch ke kaam ke liye in mein se kuch nahi chahiye. Hal: Deny karo, Settings → Permissions mein jaanch karo, achha vikalp dhoondho.
3. Bank ke AI ne Chacha ji ka loan reject kiya, par kaaran koi nahi batata.
Category: Transparency (aur Accountability). Principle: Explicability. Kyun: “black box” faisla; insaan ko samajhne aur shikayat karne ka haq hai. Hal: policy mein kaaran batana aur shikayat ka raasta zaroori.
4. Dadi ji ko “beta” ki awaaz mein video call aayi, paise maange. Baad mein pata chala deepfake tha.
Category: Safety & Security (aur Social Impact). Principle: Non-maleficence. Hal: asli number par call karke confirm, family ka “code word”, fraud par 1930 ya cybercrime.gov.in.
5. Gaon ke school mein na internet hai, na device; shehar ke bachche AI tutor se padh rahe hain.
Topic: AI Access, Digital Divide. Principle: Inclusivity. Hal: sasta internet, school mein shared devices, apni bhasha mein tools (jaise Bhashini).
6. Self-driving car ke accident ke baad company kehti hai “galti AI ki thi, hamari nahi”.
Category: Accountability. Kyun galat: AI ek tool hai; usse banane aur chalane wale zimmedar hain. Hal: policy mein pehle se tay ho ki galti par zimmedari kiski aur nuksaan ki bharpaai kaise.
⚖️ Difference Table: Sabse Confuse Hone Wali Jodiyan
| Jodi | Pehla | Doosra |
|---|---|---|
| Morals vs Ethics | Morals: insaan ke apne vishwas; aksar likhe nahi hote | Ethics: group/profession ke saajhe, aksar likhe niyam |
| Ethics vs Law | Ethics todne par samaaj/group ki naraazgi | Law sarkar banati hai; todne par saza |
| Technical vs Ethical sawaal | “Kya hum yeh bana sakte hain?” | “Kya hamein yeh banana chahiye?” |
| Personal vs Sensitive data | Naam, phone, address, photo | Password, OTP, bank, health, biometric, Aadhaar (zyada khatra) |
| Consent vs Purpose Limitation | Consent: data lene se pehle ijazat | Purpose Limitation: jis kaam ke liye liya, sirf usi mein use |
| Data Minimization vs Storage Limitation | Kam se kam data lo | Kam se kam samay tak rakho, phir mita do |
| Transparency vs Accountability | Batana ki AI hai aur kaise faisla hua | Galti par zimmedar kaun aur sudhaar kaise |
| Principle vs Policy | Principle: mool vichaar, disha (jad) | Policy: likhe niyam, kya karna hai (tana) |
| Beneficence vs Non-maleficence | Bhalaai karo | Nuksaan mat karo |
| High risk vs Unacceptable risk | Ijazat hai, par sakht niyam ke saath | Poori tarah mana |
🚫 Misconceptions: Log aisa sochte hain… lekin
- “AI galti kare toh AI zimmedar hai.” Galat. AI ek tool hai; zimmedari developer, company, sarkar aur user ki.
- “Morals aur ethics bilkul ek hi cheez hain.” Galat. Morals vyakti ke apne; ethics group ke saajhe niyam. Kuch hissa milta hai, par fark hai.
- “Jo cheez legal hai, woh hamesha ethical bhi hai.” Galat. Kanoon har situation ko cover nahi karta; kai baatein legal hokar bhi galat ho sakti hain.
- “AI ke paas data hai, toh woh hamesha nishpaksh (fair) hoga.” Galat. Data mein bhedbhaav ho toh AI bhi wahi seekhta hai.
- “Free app hai, toh koi keemat nahi.” Galat. Kabhi-kabhi keemat aapka data hota hai.
- “Ek baar permission di toh wapas nahi le sakte.” Galat. Phone ki Settings → Permissions mein kabhi bhi badal sakte ho.
- “Ethics sirf AI banane walon ka kaam hai.” Galat. Aap user bhi ho: AI ka sahi use, deepfake na banana, data surakshit rakhna aapki bhi zimmedari.
- “Ek samasya ki sirf ek hi category hoti hai.” Galat. Deepfake jaisi samasya kai categories mein aati hai (Safety, Privacy, Social Impact).
🔤 Key Terms (Exam mein English shabd hi likhna)
❓ FAQs (Exam se pehle wale sawaal)
1. Ethics ke sawaal mein “sahi jawaab” kaise tay hota hai? Sabki raay alag ho sakti hai.
HOTS ya raay wale sawaal mein ek hi sahi jawaab nahi hota; marks kaaran ke milte hain. Apni raay ke saath category, principle aur hal ka naam likho. Par definition, lists aur facts (jaise DPDP Act 2023) ke jawaab pakke hote hain.
2. Lambi lists (6 categories, 6 niyam, 8 hisse, 7 principles) kaise yaad karein?
Har list ko ek kahani se jodo aur trick banao: principles = SEI-PTA-H. Data ke niyam ko ek order mein socho: pehle ijazat (Consent) → kis kaam ke liye (Purpose) → kitna (Minimization) → surakshit (Security) → kab tak (Storage) → aapke haq (Rights).
3. Category aur principle mein kya fark hai?
Category batati hai samasya kis prakar ki hai (jaise Bias). Principle batata hai kaunsa mool vichaar toota ya kis par chalna chahiye (jaise Equality). Case mein dono likho.
4. Real examples ke saal yaad rakhne zaroori hain?
Mukhya saal yaad rakho (Diagram 3): Tay 2016, Amazon hiring AI ki khabar 2018, Gender Shades 2018, OECD 2019, NITI Aayog Feb 2021, UNESCO Nov 2021, DPDP Act 2023, EU AI Act 1 Aug 2024, New Delhi summit Feb 2026. Long answer mein ek-do example saal ke saath likhne se jawaab mazboot hota hai.
5. Agar paper mein koi naya scenario aaye jo kitaab mein nahi tha?
Ghabrao mat. Diagram 2 ke 5 steps lagao: samasya → kise asar → 5 sawaal check → kaunsa principle → hal. Yeh har naye case par kaam karta hai.
🛠️ Python Mini Project: “AI Ethics Category Finder”
Yeh chhota program har situation mein kuch pehchaan ke shabd dhoondhta hai aur batata hai ki samasya kis category ki hai. Asli AI bhi aise hi (bahut bade paimaane par) pattern dhoondhta hai, isliye iski kamzoriyan bhi wahi hain!
CharuMam AI Tutor kholo → 🐍 Python → Practice Karo tab → neeche ka code paste karo → Run. Har line ke saath comment (#) mein uska matlab likha hai.
# AI Ethics Category Finder: har situation ki category pehchano
categories = { # dictionary: category -> pehchaan ke shabd
"Bias & Fairness": ["bhedbhaav", "sirf ladke", "rang"],
"Privacy & Data": ["contacts", "location", "data"],
"Transparency": ["kyun", "bataya nahi"],
"Safety & Security": ["hack", "deepfake", "dhokha"],
}
situations = [ # list: test ke liye 4 situations
"Game app ne bina wajah contacts aur location maangi",
"Hiring AI ne sirf ladke chune",
"Loan reject hua par bank ne kyun bataya nahi",
"Dadi ko deepfake video call se dhokha",
]
ginti = 0 # kitni situations ki category mili
for s in situations: # har situation ek-ek karke
mili = "Pata nahi (insaan se poochho)" # pehle maan lo category nahi mili
for cat in categories: # har category ko check karo
for shabd in categories[cat]: # us category ke har shabd ko
if shabd in s: # kya shabd situation mein hai?
mili = cat # haan: yahi category
if mili != "Pata nahi (insaan se poochho)": # category mil gayi?
ginti = ginti + 1 # toh ginti badhao
print(s, "->", mili) # situation aur uski category
print("Category mili:", ginti, "/", len(situations))
Output:
Game app ne bina wajah contacts aur location maangi -> Privacy & Data Hiring AI ne sirf ladke chune -> Bias & Fairness Loan reject hua par bank ne kyun bataya nahi -> Transparency Dadi ko deepfake video call se dhokha -> Safety & Security Category mili: 4 / 4
Code kaam kyun karta hai?
- categories = { … } ek dictionary hai: har category (key) ke saath uske pehchaan ke shabdon ki list (value). Jaise dictionary mein shabd ke aage uska matlab.
- situations = [ … ] ek list hai jismein 4 situations hain. len(situations) = 4.
- Bahar wala for s in situations har situation ko baari-baari leta hai. Andar wale do loops har category ke har shabd ko check karte hain. (Loops ko Computer Ch 6 mein detail mein padhenge.)
- if shabd in s: in check karta hai ki shabd situation ke text ke andar hai ya nahi. Jaise “contacts” pehli situation mein hai, isliye mili = “Privacy & Data”.
- Shuru mein mili mein “Pata nahi (insaan se poochho)” rakha, taaki koi shabd na mile toh program galat category na thope, balki insaan se poochhe. Yahi human oversight hai!
- Agar category mili toh ginti = ginti + 1. Chaaron mili, isliye aakhir mein 4 / 4.
Socho: “Teacher ne bina kaaran bataye marks kaate” likho toh? Isme “kyun” ya “bataya nahi” shabd nahi hai, toh program “Pata nahi” kahega, jabki yeh Transparency ki samasya hai. Shabd-pattern wala AI matlab nahi samajhta: yahi is chapter ka Deep Concept hai.
🧩 Try It Yourself
- Program badho: categories mein “Accountability”: [“zimmedar”, “galti kiski”] jodo, aur situations mein “Accident ke baad koi zimmedar nahi bana” jodo. Run karke dekho: ab output mein kitne / kitne aaya? (Jawaab: 5 / 5)
- Ek-page Chapter Map: Diagram 1 jaisa mind map kaagaz par banao, aur har dabbe mein apne ghar ya school ka ek example likho.
- Galti ki diary: test solve karne ke baad 3 column banao: Sawaal no. | Topic | Galti kyun hui. Jis topic mein sabse zyada galti, uska post (upar Basics mein link) dobara padho.
🛡️ Good Habits: Poore Chapter ki Checklist
🧠 Think Like a Programmer: Ethics Check ko Program Banana
Problem: Ek AI app ke baare mein 5 haan/naa sawaal poochh kar batana ki woh “Theek”, “Sudhaar chahiye” ya “Khatarnaak” hai.
Steps: (1) 5 sawaalon ke jawaab ek list mein rakho (True = theek). (2) True gino. (3) 5 = Theek, 3–4 = Sudhaar chahiye, 0–2 = Khatarnaak.
# 5 Sawaal Ethics Check: ek AI app ko jaancho
app = "Smart Padhai" # jis app ko jaanchna hai
jawaab = [True, True, False, True, True] # 5 sawaalon ke jawaab (True = theek)
haan = 0 # True ki ginti, shuru mein zero
for j in jawaab: # har jawaab ko ek-ek karke dekho
if j == True: # agar jawaab theek hai
haan = haan + 1 # toh ginti badhao
print("App:", app) # app ka naam
print("Theek jawaab:", haan, "/", len(jawaab)) # kitne mein se kitne theek
if haan == 5: # paanchon theek?
print("Nateeja: Theek")
elif haan >= 3: # 3 ya 4 theek?
print("Nateeja: Sudhaar chahiye")
else: # 0, 1 ya 2 theek
print("Nateeja: Khatarnaak")
Output:
App: Smart Padhai Theek jawaab: 4 / 5 Nateeja: Sudhaar chahiye
Test: jawaab = [True, True, False, True, False] → 3 True → “Sudhaar chahiye” ✅ (3, 3 se bada ya barabar hai).
Fix: Pehle humne likha tha if haan > 3. Tab 3 True wala app “Khatarnaak” ban jaata, jo galat hai. Isliye >= kiya. Ek aur sudhaar socho: agar “Ijazat/Privacy” wala jawaab hi False hai toh baaki sab True hone par bhi app “Theek” nahi hona chahiye. Programmer aise hi niyam (policy) ko code mein sudhaarte rehte hain. 🔧
🌍 Real Life Mein AI Ethics
- Mobile: app permissions, face unlock (biometric data), recommendation feed jo aapko ghanton tak rokta hai.
- School: AI attendance, AI tutor, online exam proctoring: barabari aur privacy dono ka sawaal.
- Dukaan aur bank: UPI fraud, loan ke AI faisle, deepfake calls se dhokha.
- Cricket: AI se player selection ya ball-tracking: faisla kaise hua yeh batana (transparency) zaroori, warna bharosa toot jaata hai.
🔁 Recap: Poora AI Chapter 3 Ek Nazar Mein
- AI Ethics: AI ko sahi tareeke se banane, train karne, use karne ke siddhant. “Bana sakte hain?” (technical) vs “Banana chahiye?” (ethical). Zaroori kyunki: speed-scale, data se bias, sahi-galat ki samajh nahi, black box, personal data.
- 5 Sawaal Ethics Check: Fayda/Nuksaan, Barabari, Ijazat/Privacy, Saaf-safaai, Zimmedari. AI lakshya poora karta hai, matlab nahi samajhta.
- Morals (vyakti ke) · Ethics (group ke, likhe) · Law (sarkar ke, saza). AI ke apne morals nahi hote.
- 6 categories: Bias & Fairness · Privacy & Data · Transparency · Accountability · Safety & Security · Social Impact.
- AI Access + Digital Divide; access ki chintaayein: barabari, kharcha, bhasha, accessibility, galat istemaal, umr, nirbharta, taaqat kuch haathon mein.
- Data ke 6 niyam: Consent · Purpose Limitation · Data Minimization · Security · Storage Limitation · Rights. DPDP Act, 2023; 18 se kam: maata-pita ki verifiable consent.
- AI Policy ke 8 hisse; kisse: Tay (2016), Amazon hiring AI (2018), Gender Shades (2018), Netherlands (Jan 2021); achhe examples: Bhashini, EU AI Act, NITI Aayog. Risk-based: Unacceptable → High → Limited → Minimal.
- Principles: NITI Aayog ke 7 (SEI-PTA-H), OECD (2019, update 2024), UNESCO (Nov 2021, 193 desh); 5 classic: Beneficence, Non-maleficence, Autonomy, Justice, Explicability. Ethics by Design.
- Future: deepfakes, naukriyaan, Green AI, bachche aur AI; New Delhi summit (Feb 2026): “Sarvajan Hitaya, Sarvajan Sukhaya”; People, Planet, Progress.
✍️ Warm-up 1: Fill in the Blanks
- Vyakti ke apne sahi-galat ke vishwas ______ kehlate hain.
- Technology wale aur bina technology wale logon ke beech ki khaai ko ______ kehte hain.
- Sirf utna data lena jitna kaam ke liye zaroori ho, ______ kehlata hai.
- Bharat ka digital personal data kanoon ______ Act, 2023 hai.
- NITI Aayog ke 7 principles yaad rakhne ki trick ______ hai.
- “Nuksaan mat karo” wale siddhant ko ______ kehte hain.
Answers
1) Morals · 2) Digital Divide · 3) Data Minimization · 4) DPDP (Digital Personal Data Protection) · 5) SEI-PTA-H · 6) Non-maleficence
✅ Warm-up 2: True / False
- AI ke apne morals hote hain kyunki woh insaan jaisa sochta hai.
- Ek hi samasya (jaise deepfake) ek se zyada ethics categories mein aa sakti hai.
- Phone ki di hui permission baad mein wapas nahi li ja sakti.
- Risk-based approach mein “Unacceptable risk” wale AI poori tarah mana hote hain.
- Policy pehle banti hai, principles baad mein policy se nikalte hain.
- Gender Shades study mein face AI ki galti saanwli mahilaon par sabse zyada thi.
Answers (kyun ke saath)
1) False: AI ke paas zameer aur bhaavna nahi; ethics insaan data, niyam aur nigraani se daalte hain.
2) True: deepfake Safety, Privacy aur Social Impact teeno mein aata hai.
3) False: Settings → Permissions se kabhi bhi badal sakte hain.
4) True: jaise logon ko chupke se nuksaan pahunchane wala AI.
5) False: ulta; principle jad hai, usse policy (tana) aur phir niyam (pattiyaan) bante hain.
6) True: lagbhag 35% tak, jabki gore purushon par 1% se kam.
📝 Model Test Paper: AI Chapter 3 (AI Ethics and Morality)
Section A: MCQ (10 × 1 = 10) · Section B: Assertion–Reason (4 × 1 = 4) · Section C: Very Short Answer (6 × 1 = 6) · Section D: Short Answer (4 × 2 = 8) · Section E: Case-Based (4 × 1 = 4) · Section F: Long Answer (2 × 4 = 8)
Niyam: pehle copy mein jawaab likho, phir “Answer” kholo. Imaandaari se marks gino! (Yeh bhi ethics hai 😊)
🔘 Section A: MCQs 10 marks
Q1. “Kya hamein yeh AI banana chahiye?” kis prakar ka sawaal hai? (a) Technical (b) Ethical (c) Mathematical (d) Legal
✅ (b) Ethical: “chahiye” sahi-galat ki baat hai.
(a) Technical sawaal hota “bana sakte hain?”.
(c) Isme koi ganit nahi.
(d) Legal sawaal hota “kanoon ijazat deta hai?”.
Q2. Kisi group ya profession ke saajhe, aksar likhe hue niyam kya kehlate hain? (a) Morals (b) Bhaavna (c) Ethics (d) Algorithm
✅ (c) Ethics: jaise doctors ka code of conduct.
(a) Morals vyakti ke apne vishwas hote hain.
(b) Bhaavna feeling hai, niyam nahi.
(d) Algorithm kaam karne ke steps hain, sahi-galat ke niyam nahi.
Q3. Hiring AI ne purane data se seekh kar ladkiyon ke resume ko kam number diye. Yeh kis category ki samasya hai? (a) Bias & Fairness (b) Storage Limitation (c) Green AI (d) Digital Divide
✅ (a) Bias & Fairness: data ka bhedbhaav AI ne seekh liya.
(b) Storage Limitation data kab tak rakhein, uska niyam hai.
(c) Green AI bijli-paani bachane ki baat hai.
(d) Digital Divide technology tak pahunch ki khaai hai.
Q4. Torch app ka contacts maangna kis data niyam ko todta hai? (a) Rights (b) Data Minimization (c) Explicability (d) Autonomy
✅ (b) Data Minimization: torch ko contacts ki zaroorat hi nahi.
(a) Rights aapke data par aapke haq hain (dekhna, sudharna, mitwana); yahan mukhya galti zyada data lena hai.
(c) Explicability faisla samjhane ka siddhant hai, data niyam nahi.
(d) Autonomy insaan ki apni marzi ka siddhant hai.
Q5. Bharat ka DPDP Act kis saal aaya? (a) 2016 (b) 2019 (c) 2021 (d) 2023
✅ (d) 2023: Digital Personal Data Protection Act, 2023.
(a) 2016 mein Tay chatbot ka kissa hua tha.
(b) 2019 mein OECD AI Principles aaye.
(c) 2021 mein NITI Aayog aur UNESCO ke principles aaye.
Q6. Inmein se kaun NITI Aayog ke 7 principles mein se nahi hai? (a) Equality (b) Transparency (c) Profit First (d) Accountability
✅ (c) Profit First: SEI-PTA-H mein munafa koi principle nahi.
(a) Equality = E.
(b) Transparency = T.
(d) Accountability = A.
Q7. “AI ka faisla insaan ko samjhaya ja sake” kis siddhant ka matlab hai? (a) Explicability (b) Beneficence (c) Justice (d) Non-maleficence
✅ (a) Explicability: samjha sakna.
(b) Beneficence = bhalaai karna.
(c) Justice = nyay, barabari.
(d) Non-maleficence = nuksaan na karna.
Q8. Risk-based approach mein chatbot ka “main AI hoon” batana zaroori hona kis level mein aata hai? (a) Unacceptable (b) High (c) Limited (d) Minimal
✅ (c) Limited: is level mein mukhya niyam hai batana (transparency).
(a) Unacceptable poori tarah mana hota hai.
(b) High mein sakht niyam (testing, nigraani) hote hain.
(d) Minimal jaise spam filter, khaas niyam nahi.
Q9. Microsoft ka Tay chatbot (2016) jaldi band kyun karna pada? (a) Bijli zyada khaata tha (b) Logon se galat baatein seekh kar aapattijanak post karne laga (c) Uska data chori ho gaya (d) Woh sirf Hindi bolta tha
✅ (b): testing aur nigraani ki kami; ek din ke andar band.
(a) Bijli ka masla nahi tha.
(c) Data chori ki baat nahi thi.
(d) Galat; bhasha ka masla nahi tha.
Q10. AI policy mein “AI ke faisle par insaan nazar rakhe aur zaroorat par badal sake” kaunsa hissa hai? (a) Uddeshya (b) Human Oversight (c) Storage Limitation (d) Digital Divide
✅ (b) Human Oversight: insaan ki nigraani.
(a) Uddeshya batata hai AI kis kaam ke liye hai.
(c) Storage Limitation data ka niyam hai, policy ka yeh hissa nahi.
(d) Digital Divide access ki khaai hai.
🔗 Section B: Assertion–Reason 4 marks
Options: (a) A aur R dono sahi, R, A ko samjhata hai · (b) A aur R dono sahi, par R, A ko nahi samjhata · (c) A sahi, R galat · (d) A galat, R sahi
1. A: AI galat data se bhedbhaav seekh sakta hai. R: AI data mein chhupe pattern se seekhta hai.
✅ (a): pattern se seekhta hai, isliye data ka bhedbhaav bhi pattern ban kar seekh leta hai.
2. A: AI galti kare toh uski zimmedari AI ki hoti hai. R: AI ke paas zameer aur bhaavna nahi hoti.
✅ (d): A galat (zimmedari insaan ki); R sahi. Ulta, R hi batata hai ki AI zimmedar nahi ho sakta.
3. A: DPDP Act, 2023 bachchon ke data ke liye maata-pita ki verifiable consent maangta hai. R: EU AI Act 1 August 2024 ko laagu hua.
✅ (b): dono baatein sahi, par EU ka kanoon Bharat ke DPDP niyam ko nahi samjhata.
4. A: Digital Divide AI ke faayde sabko barabar milne se rokta hai. R: AI tools use karne ke liye sirf bijli chahiye.
✅ (c): A sahi; R galat, kyunki device, internet, bhasha, gyaan aur kabhi paise bhi chahiye.
📝 Section C: Very Short Answer 6 × 1
1. AI Ethics ki ek line ki paribhasha likho.
AI ko sahi tareeke se banane, train karne aur use karne ke siddhant, taaki kisi ka nuksaan na ho aur faayda sabko mile.
2. Do sensitive personal data ke udaharan do.
Koi do: password, OTP, bank details, health record, biometric (fingerprint/face), Aadhaar number.
3. AI ethics ki koi do categories ke naam likho.
Koi do: Bias & Fairness, Privacy & Data, Transparency, Accountability, Safety & Security, Social Impact.
4. UNESCO ki AI Ethics Recommendation kab apnayi gayi?
November 2021 (193 sadasya deshon ne).
5. Deepfake kya hai?
AI se banaya gaya nakli video, photo ya awaaz jo asli jaisa lage.
6. Online fraud ki shikayat ke liye Bharat ka helpline number?
1930 (website: cybercrime.gov.in).
📝 Section D: Short Answer 4 × 2
1. Morals aur Ethics mein do antar likho.
(i) Morals vyakti ke apne vishwas hain (parivaar, sanskriti, anubhav se); Ethics kisi group ya profession ke saajhe niyam hain.
(ii) Morals aam taur par likhe nahi hote aur vyakti-vyakti mein alag ho sakte hain; Ethics aksar likhe hote hain (code of conduct) aur group ke sab sadasyon par ek jaise laagu hote hain.
2. AI ethics zaroori kyun hai? Koi do kaaran.
Koi do: (i) AI bahut tez aur bade paimaane par faisle karta hai, toh ek galti laakhon par asar daalti hai; (ii) AI data se seekhta hai, toh data ka bhedbhaav bhi seekh leta hai; (iii) AI ko sahi-galat ki samajh nahi; (iv) kai AI black box hain; (v) AI personal data use karta hai.
3. Consent aur Purpose Limitation ko ek-ek udaharan ke saath samjhao.
Consent: data lene se pehle saaf shabdon mein ijazat lena. Jaise school app photo lene se pehle parents se poochhe.
Purpose Limitation: data sirf usi kaam mein use karna jiske liye liya. Jaise attendance ke liye li gayi photo ko advertisement mein use na karna.
4. Risk-based approach kya hai? Chaar level likho.
AI ke khatre ke hisaab se niyam: jitna bada khatra, utne sakht niyam. Levels: Unacceptable (mana) → High (sakht niyam) → Limited (batana zaroori) → Minimal (khaas niyam nahi).
📰 Section E: Case-Based Question 4 × 1
(i) Location aur address lena kis data niyam ke khilaaf hai?
Data Minimization (padhai ke liye inki zaroorat nahi). Purpose Limitation bhi likh sakte ho.
(ii) Smartphone na hone wale bachchon ki samasya ko kya kehte hain?
Digital Divide (AI Access ki samasya); principle: Inclusivity.
(iii) “Humein nahi pata AI ne aisa kyun kiya” kaunsi do categories ki kami dikhata hai?
Transparency (faisla samjha nahi sakte, black box) aur Accountability (“galti app ki” kehna zimmedari se bachna hai).
(iv) Data kab tak rakhein, na batana kis niyam ki kami hai? Ek sudhaar bhi likho.
Storage Limitation. Sudhaar: policy mein likho ki data kitne samay tak rahega, phir apne aap mitega; aur parents ko Rights do ki woh data dekh ya mitwa sakein.
📝 Section F: Long Answer 2 × 4
1. Ek achhi AI policy ke koi chaar hisse samjhao. Ek real example likho jahan policy ki kami se nuksaan hua.
Koi chaar (har ek ki ek line):
1. Uddeshya: AI kis kaam ke liye hai, kis ke liye nahi.
2. Fairness: data aur nateeje mein bhedbhaav ki jaanch.
3. Privacy: consent, kam data, surakshit rakhna.
4. Transparency: logon ko batana ki AI hai aur faisla kaise hua.
5. Human Oversight: zaroori faislon par insaan ki nazar.
6. Zimmedari & Shikayat: galti par kaun zimmedar, shikayat kahan.
7. Safety, Testing & Review: launch se pehle testing, baad mein review.
Example (koi ek): Tay chatbot (2016): testing aur nigraani ki kami, galat baatein seekh kar ek din mein band; ya Amazon hiring AI (khabar 2018): data ka bhedbhaav seekha, project band; ya Netherlands childcare kissa: parivaar galat dhokhebaaz maane gaye, sarkar ne Jan 2021 mein istifa diya.
2. 5 classic AI ethics siddhant likho aur har ek ka ek line mein matlab ya udaharan do. AI ethics aap jaise students ke liye kyun zaroori hai?
1. Beneficence: bhalaai karo (jaise AI se doctor ko jaldi bimari pakadne mein madad).
2. Non-maleficence: nuksaan mat karo (deepfake se kisi ko badnaam na karo).
3. Autonomy: insaan ki marzi (aakhri faisla insaan ka, AI sirf salah de).
4. Justice: nyay, barabari (sab bachchon ke chehre attendance AI pehchane).
5. Explicability: faisla samjha sakna (loan reject ho toh kaaran batao).
Students ke liye kyun: aap aaj AI ke user ho (data, deepfake, padhai mein AI) aur kal ke banane wale. Ethics se bharosa, adhikaar aur suraksha bani rehti hai.
Apna score: 36–40: Shandaar! 🏆 · 28–35: Bahut achha, galtiyon wale post ek baar padho · 20–27: Thodi aur mehnat, Recap aur Difference Table dobara · 20 se kam: Chapter ke posts ek-ek karke dobara, phir yeh test phir se 💪
🐍 Output Pehchano / Galti Dhoondho (Python)
1. Iska output kya hoga?
scores = [4, 2, 5]
total = 0
for x in scores:
total = total + x
print(total, round(total / len(scores), 1))
Output: 11 3.7
Kyun: 4 + 2 + 5 = 11; len(scores) = 3; 11 / 3 = 3.666…, round(…, 1) = 3.7.
2. Is code mein galti dhoondho:
haan = 4
if haan >= 3
print("Sudhaar chahiye")
Galti: if wali line ke aakhir mein colon (:) nahi hai. Sahi: if haan >= 3:. Theek karne par output: Sudhaar chahiye (4, 3 se bada hai).
3. Output kya hoga?
s = "App ne bina poochhe location li"
print("location" in s, "contacts" in s)
Output: True False
Kyun: “location” shabd text mein hai, “contacts” nahi. Category Finder isi in se kaam karta hai.
🚀 HOTS
1. Ek city police “galiyon mein face recognition camera” lagana chahti hai taaki chor pakde jaayein. Ethics ke do faayde aur do khatre batao, aur ek beech ka raasta sujhao.
Faayde: (i) apradh jaldi pakda ja sakta hai (Beneficence); (ii) logon ko suraksha ka ehsaas.
Khatre: (i) Privacy: har aam aadmi par nazar; (ii) Bias: kuch chehron par zyada galti (Gender Shades), begunaah pakde ja sakte hain.
Beech ka raasta: sirf zaroori jagah, saaf likhi policy, data kam samay tak, har match par insaan ki jaanch (human oversight), shikayat ka raasta, aur samay-samay par bias testing. Yeh “Privacy vs Safety” takraav ka hal hai: kam nuksaan, khule mein faisla, dobara review.
2. Priya kehti hai, “AI homework app ne jawaab de diya, toh copy karna galat kaise? App toh legal hai!” Usse kaise samjhaoge?
Legal hona ≠ ethical hona. App legal hai, par bina samjhe copy karna imaandaari ke khilaaf hai (school ke ethics), aur Priya ka apna nuksaan (seekhna ruk gaya, Autonomy: apni soch kamzor). AI ka jawaab galat bhi ho sakta hai. Sahi tareeka: AI se samjho, phir khud likho, aur zaroorat ho toh teacher ko batao ki AI se madad li.
3. Category Finder program mein “Mera data bina poochhe bech diya” likhne par output kya aayega? Kya woh sahi hai? Program ko behtar kaise banaoge?
Situation mein “data” shabd hai, toh output Privacy & Data aayega: sahi. Par “Meri jaankari bina poochhe bech di” likhte toh koi shabd nahi milta aur “Pata nahi” aata. Sudhaar: zyada shabd (jaankari, bech, photo), ek situation ke liye kai categories dikhana, aur shak mein insaan se poochhna. Asli AI mein yahi kaam bahut bade data se training karke hota hai, aur tab bhi testing aur nigraani zaroori hai.
🏆 Cyber Olympiad Corner
1. Ek face AI ne 1,000 saanwli mahilaon mein se 350 ko galat pehchana, aur 1,000 gore purushon mein se 8 ko. Dono ki galti dar (error rate) mein kitna fark hai? (a) 34.2% (b) 35% (c) 0.8% (d) 342%
✅ (a) 34.2%: 350/1000 = 35%, 8/1000 = 0.8%; 35 − 0.8 = 34.2 percentage points.
(b) Sirf pehli dar likhi.
(c) Sirf doosri dar likhi.
(d) Percent ka hisaab galat (342 ginti ka fark hai, percent nahi).
2. Ek app 6 data niyamon mein se Consent, Security aur Rights maanta hai, par baaki tod deta hai. Kaunse niyam toote? (a) Consent, Purpose Limitation, Rights (b) Purpose Limitation, Data Minimization, Storage Limitation (c) Security, Storage Limitation, Rights (d) Sirf Data Minimization
✅ (b): 6 − 3 = 3 bache: Purpose Limitation, Data Minimization, Storage Limitation.
(a) Consent aur Rights toh maane gaye the.
(c) Security aur Rights maane gaye the.
(d) Teen niyam toote, ek nahi.
3. “SEI-PTA-H” mein “I” kis principle ke liye hai? (a) Intelligence (b) Inclusivity & Non-discrimination (c) Innovation (d) Internet
✅ (b) Inclusivity & Non-discrimination.
(a), (c), (d) NITI Aayog ke 7 principles mein nahi hain.
4. Kaunsa jodaa galat hai? (a) OECD AI Principles: 2019 (b) UNESCO Recommendation: Nov 2021 (c) EU AI Act laagu: 1 Aug 2024 (d) NITI Aayog Responsible AI: 2016
✅ (d): NITI Aayog ke Responsible AI principles Feb 2021 mein aaye; 2016 Tay chatbot ka saal hai.
(a), (b), (c) teeno sahi jode hain.
5. Category Finder mein agar mili = cat ki jagah pass likh diya jaaye, toh aakhri line mein kya chhapega? (a) Category mili: 4 / 4 (b) Category mili: 0 / 4 (c) Category mili: 1 / 4 (d) Error aayega
✅ (b) Category mili: 0 / 4: pass kuch nahi karta, toh mili hamesha “Pata nahi…” rahega aur ginti kabhi nahi badhegi.
(a) Tab hota jab mili badalta.
(c) Koi bhi situation ginti nahi badhayegi.
(d) Error nahi aayega, kyunki if ke neeche pass hai; agar line poori hata dete aur kuch na likhte, tab IndentationError aata.
📚 Aur Padho (External Reference)
- UNESCO: Recommendation on the Ethics of Artificial Intelligence
- OECD AI Principles
- MeitY (Bharat sarkar): DPDP Act aur IndiaAI
- National Cyber Crime Reporting Portal: helpline 1930
- Python docs: Dictionaries: Category Finder ka dictionary samajhne ke liye
💖 CharuMam Motivation
Test ke baad Rohan ne maana: “Ma’am, case-based sawaal mein maine 5 steps lagaye: samasya, kise asar, 5 sawaal, principle, hal. Poore number aaye! Ethics ‘kuch bhi’ nahi tha, isme bhi soch ka tareeka hai.” 🎉
Beta, AI kitna bhi smart ho jaaye, sahi-galat ka faisla hamesha insaan ke haath mein hai. Aur yeh faisla bade mauke par nahi, chhoti aadaton mein banta hai: test mein imaandaari se marks ginna, kisi ki photo bina poochhe na bhejna, galti hone par “meri zimmedari” kehna. Mehnat se seekho, sabr se socho, aur apni soch ko focus mein rakho. Aap sirf AI ke user nahi, kal ke zimmedar banane wale ho. 🌱🤖
— CharuMam
Koi doubt? CharuMam AI Tutor mein 💻 Computer subject chuno aur poochho! 💬 · Saare chapters (Computer & AI)
Yeh content CharuMam ki original rachna hai. Bina permission copy, PDF ya re-upload karna mana hai.
charumam.com ek independent website hai; iska kisi publisher, Google, Adobe ya OpenAI se koi sambandh nahi hai.
Share karna ho toh link share karein 🙏
