AI Fluency: Ek Crash Course

4 salahiyatein. AI ke saath kaam karne ke 3 tareeqe. Ek amali aadat: faisla karein, wazahat karein, jaanchein, aur zimmedari lein.

Sochiye ke ek ba-salahiyat nayi colleague aapki team mein shamil hoti hai.

Uski pehli subah aap kehte hain:

"AI agents par course ka outline taiyar karein."

Kuch ghanton baad woh aapko ek polished outline deti hai. Magar masla hai. Woh PhD researchers ke liye likha gaya hai. Usmein poore semester ka waqt farz kiya gaya hai. Aur hands-on practice na hone ke barabar hai.

Kya woh na-ahl thi? Nahin.

Asal masla yeh hai ke aapne use kaafi kuch nahin bataya. Aapne audience, maujood waqt, teaching style, ya yeh nahin samjhaya ke aakhir mein students kya kar sakne chahiye.

AI ke saath kaam bhi aisa hi hai, lekin ek ahem farq ke saath.

Nayi colleague seekhti hai. Ek baar batayein ke aapke students beginners hain, to use agle mahine bhi yaad rahega. AI aise kaam nahin karti. Woh ek conversation se doosri mein kuch nahin le jati, is liye har chat aisi colleague ke saath khulti hai jo aapse kabhi nahin mili. Jo baat aap kisi shakhs ko ek baar batate, woh AI ko har baar dobara batayein ya aisi jagah rakhein jahan se AI use khud parh le. Yeh shikayat wali kharabi nahin. Yeh kaam ki ek shart hai jiske mutabiq planning karni hoti hai, aur yeh planning neeche di gayi chaar skills mein se ek hai.

Jab collaboration kamzor ho to powerful AI bhi ghalat nateeja de sakti hai. Chand hoshiyar prompts jaanna kaafi nahin. Aapko jaanna hoga ke AI ko kya dena hai, use kaise guide karna hai, uske kaam ko kaise parakhna hai, aur kab us par bharosa ya use nahin karna.

Isi ka naam AI fluency hai.

Yeh course AI Fluency Framework sikhata hai, jise Professor Rick Dakan aur Professor Joseph Feller ne banaya aur Anthropic ke saath taiyar kiye gaye courses mein parhaya. Framework mein insaan ki chaar salahiyatein hain jinhein 4Ds kaha jata hai:

1. Delegation: faisla karein ke AI ko kya karna chahiye.
2. Description: wazeh karein ke aapko kya chahiye.
3. Discernment: AI ke diye hue kaam ko parakhein.
4. Diligence: AI ko zimmedari se istemaal karein aur nateeje ki zimmedari lein.

Aap yehi chaar skills istemaal karenge, chahe assistant se chat kar rahe hon, AI agent ke saath coding kar rahe hon, ya doosron ke liye kaam karne wala Digital FTE bana rahe hon.

Reading time: taqreeban 30 minutes, aur practice prompts aur self-check ke liye mazeed 15-20 minutes.

Ek minute mein poora course

Agar sirf chaar sawal yaad rakhein, to yeh rakhein:

Salahiyat	Seedha sawal
Delegation	AI ko kya karna chahiye, aur kya mere paas rehna chahiye?
Description	Kaam achi tarah karne ke liye AI ko kya jaanna zaroori hai?
Discernment	Kya nateeja waqai acha aur bharose ke qabil hai?
Diligence	Kya AI ka yeh istemaal zimmedarana hai, aur kya main nateeja apna sakta hun?

Baqi course mein aap in chaar sawalon ke ache jawab dena seekhenge.

Foundations mein iski jagah

Is course se pehle What AI Actually Is parhein. Woh course machine samjhata hai. Yeh course samjhata hai ke aapko machine ke saath kaise kaam karna chahiye. Iske baad aane wala AI Prompting in 2026 amali techniques sikhata hai.

Ek mufeed tarteeb yeh hai:

Machine -> collaboration -> techniques

Topic	What AI Actually Is	Yeh course	AI Prompting in 2026
AI kaise kaam karti hai	Tafseel se	Mukhtasar yaad	Pehle se samjha hua
Baat kaise samjhani hai	Context kyun ahem hai	Description	Amali prompting techniques
Jawab kaise parakhna hai	Plausible baat ghalat kyun ho sakti hai	Discernment	Model-checking ki aadatein
AI ko kya dena hai	Jagged frontier	Delegation	Models aur tools ka intikhab
Zimmedarana istemaal	Zyada tar scope se bahar	Diligence	Tools aur permissions ka mehfooz istemaal

Models behtar hone ke saath techniques badlein gi. Yeh chaar salahiyatein derpa rehne ke liye banayi gayi hain.

Teen minute mein dekh lein

Theory se pehle woh cheez hoti hui dekhein jis ke baare mein yeh course hai.

Koi bhi AI assistant kholein. Yeh paste karein aur jawab parhein:

Write a welcome email for new members.

Jawab dhyan se parhein. Woh theek, grammatical, aur bilkul generic hoga, kyunke aapne kaam ke liye koi maloomat nahin di. Ab ek fresh chat kholein aur yeh paste karein:

Write a welcome email for new members of a small women's cycling club
in Karachi. Most are nervous beginners who have never ridden in traffic.
Warm and a bit funny, under 150 words, no exclamation marks. End by
telling them the Saturday 6am ride is slow on purpose and nobody gets
dropped.

Wohi model. Wohi task. Machine ke liye wohi teen seconds ka kaam.

Doosri email behtar hai kyunke doosri request mein woh maloomat thi jo pehli mein reh gayi: yeh log kaun hain, kis baat se darte hain, lehja kaisa ho, lambai kitni ho, aur maqsad kya hai. Koi hoshiyari nahin hui. Bas woh baat keh di gayi jo pehle nahin kahi gayi thi.

Jo do baatein aapne abhi dekhin, unhein yaad rakhein:

1. Dono results ka farq model se nahin, aapse aaya.
2. Aap sirf is liye jaan sake ke doosri email behtar hai kyunke aap cycling clubs, nervous beginners, ya Karachi ke baare mein itna jaante the ke use parakh sakein.

Pehli Description hai. Doosri Discernment. Teen minute mein chaar mein se do skills pehle hi aapke paas hain. Baqi course chaaron ko naam deta hai, dikhata hai ke har ek kahan fail hoti hai, aur batata hai ke chat window mein akele kaam karne se aage barh kar doosron ke liye AI banane par har skill kya ban jati hai.

Aakhir tak aapko kya samajh ana chahiye

Course ke aakhir tak aap yeh samjha sakne chahiye:

• AI fluency ka matlab "prompts mein acha hona" se aage kya hai;
• automation, augmentation, aur agency mein farq;
• kaise faisla karein ke kaunsa kaam aapka aur kaunsa AI ka hai;
• Description ki teen qisamain: product, process, aur performance;
• confident lagne ki wajah se maan lene ke bajaye AI output ko kaise parakhein;
• Diligence ki teen qisamain: creation, transparency, aur deployment;
• asal project par chaaron salahiyatein kaise mil kar kaam karti hain; aur
• yeh zaati skills Agent Factory engineering practices mein kaise scale hoti hain.

Hissa 1: Bari tasveer se shuru karein

1. AI tak rasaai, AI fluency nahin hai

Powerful AI tak rasaai ka matlab yeh nahin ke aap use achi tarah istemaal karna jaante hain.

Yeh parhne wale taqreeban sab logon ko wohi models dastiyab hain. Do log ek hi assistant, ek hi plan, aur ek hi subah kholte hain. Ek ko ship karne layak kaam milta hai, doosre ko polished cheez milti hai jo woh chup chap phenk deta hai. Tool mein farq nahin tha. Farq use karne ke tareeqe mein tha.

AI fluency ka matlab AI ke saath in tareeqon se kaam karna hai:

• Effective: aap goal tak pohanchte hain.
• Efficient: be-zaroorat waqt, mehnat, ya tokens zaya nahin karte.
• Ethical: AI ko insaf aur imaandari se istemaal karte hain.
• Safe: logon, privacy, security, aur ahem maloomat ko mehfooz rakhte hain.

Ghaur karein ke is definition mein kya nahin hai.

Aapko large language model train karna aana zaroori nahin. Transformers ki har tafseel samajhna zaroori nahin. "Magic prompts" ka majmua bhi zaroori nahin.

Yeh cheezein mufeed ho sakti hain, magar buniyad nahin hain.

Buniyad AI ke gird ache insani faisle karna seekhna hai.

Isi liye framework 4Ds par tawajjoh deta hai:

Delegation -> Description -> Discernment -> Diligence

Inhein yaad rakhne ka seedha tareeqa hai:

Faisla -> Wazahat -> Jaanch -> Zimmedari

Official competency names ahem hain aur poori book mein istemaal honge. Seedhe alfaaz wala version har ek ka kaam yaad rakhne ke liye hai.

Agent Factory readers ke liye yeh har cheez se pehle aata hai. Mode 1 mein aap general agents se masle hal karte hain. Mode 2 mein doosron ke liye Digital FTEs banate hain. Dono ko AI fluency chahiye.

Agar aap ek AI assistant ke saath acha kaam nahin kar sakte, to aise AI system ko design karne ke liye taiyar nahin jo sau ya hazaron users ki taraf se kaam kare.

Pichle course se teen facts yaad rakhein

What AI Actually Is ne language models ke baare mein kai ideas sikhaye. Yeh teen zehan mein rakhein:

1. Plausible hona correct hone ke barabar nahin. AI confident magar ghalat jawab de sakti hai. Ise aksar hallucination kehte hain.
2. Outputs badal sakte hain. Wohi request mukhtalif waqt par mukhtalif jawab de sakti hai.
3. AI sirf dastiyab maloomat ke saath kaam karti hai. Ismein trained knowledge, current conversation, documents, memory, search, aur doosre connected tools shamil ho sakte hain. Ahem maloomat na ho to AI andaza laga sakti hai.

Is liye AI output ko ek ba-salahiyat colleague ka kaam samjhein: mufeed aur aksar impressive, magar phir bhi review karna zaroori hai.

2. AI ke saath kaam ke teen tareeqe: automation, augmentation, aur agency

4Ds seekhne se pehle ek aur idea zaroori hai.

Insaan AI ke saath teen wasee tareeqon se kaam kar sakta hai:

1. Automation: AI aapka diya hua task karti hai.
2. Augmentation: aap aur AI task par mil kar kaam karte hain.
3. Agency: AI aapki taraf se goal ki taraf kaam karti hai.

Asal farq yeh hai ke agla qadam chunne ki kitni azadi AI ke paas hai.

Automation: "Yeh task karein"

Automation mein aap AI ko bilkul batate hain ke kaunsa task karna hai.

Misalein:

• "Is report ko paanch bullets mein summarize karein."
• "Is email ko Urdu mein translate karein."
• "Is PDF se invoice number, date, aur total nikalein."

Aap script writer ke qareeb hain. Aap task tay karte hain. AI use karti hai.

Automation tab mufeed hai jab kaam wazeh aur dohraya ja sakta ho.

Augmentation: "Sochne mein meri madad karein"

Augmentation mein aap aur AI mil kar kaam karte hain.

Misalein:

• business idea par brainstorming;
• software architecture ka review;
• lesson plan ko behtar banana;
• do strategies ka muqabla; ya
• aisa sawal explore karna jiska jawab abhi aap nahin jaante.

Yahan AI sirf instructions nahin chala rahi. Woh thinking partner jaisa kirdar ada karti hai.

Kai turns tak baat aage peeche ja sakti hai. Aap poochte hain. Woh jawab deti hai. Aap challenge karte hain. Woh revise karti hai. Dono mil kar nateeja behtar karte hain.

Agency: "Meri taraf se is goal ka peecha karein"

Agency mein aap AI ko goal aur boundaries dete hain, phir kai qadmon ka faisla use karne dete hain.

Misaal ke taur par, yeh kehne ke bajaye:

"Yeh paanch emails parhein aur summarize karein."

Aap keh sakte hain:

"Mera inbox manageable rakhein. Routine messages ka jawab dein, ahem messages flag karein, aur jis kaam par yaqeen na ho use karne se pehle mujhse poochein."

Ab AI ko faisle karne hain. Kaunsa message routine hai? Kaunsa ahem hai? Use aapse kab poochna chahiye?

Aap script writer se director ban gaye hain.

Agency ka doosra hissa beginners miss karte hain, aur isi par yeh book bani hai. Framework ki apni definition ke mutabiq insaan AI ko mustaqbil ke tasks apni taraf se azaad taur par karne ke liye configure karta hai, jin mein doosron ke liye kaam bhi shamil hai. Dobara ahista parhein. Do alfaaz asal bojh uthate hain.

Mustaqbil ka matlab hai aap kamre mein nahin honge. Aap Monday ko AI set karte hain aur woh aapke sote hue Thursday ka kaam sambhalti hai.

Doosron ke liye ka matlab hai AI jis shakhs ko serve karegi woh shayad aap na hon. Aap configure karte hain; usse aapka customer, student, ya colleague baat karta hai.

Automation aur augmentation mein aap kursi par rehte hain. Agency mein aap kursi se uth jate hain. Mode 2 ko mushkil banane wali har cheez isi fact se nikalti hai: aap har faisla supervise nahin kar sakte, is liye judgment pehle se system mein banana hota hai.

Automation aur agency ka farq khaas taur par ahem hai:

	Automation	Agency
Aap dete hain	Task ya steps	Goal aur boundaries
AI tay karti hai	Bahut kam	Kai agle qadam
Aapka kirdar	Script writer	Director
Aam failure	Ek qadam bura hua	Goal ya boundary ghalat samjhi gayi

In mein se koi mode khud-bakhud behtar nahin.

Acha AI user kaam ke mutabiq mode chunta hai.

Ek project mein teeno ho sakte hain. Aap data extraction automate kar sakte hain, exceptions par sochne ke liye augmentation use kar sakte hain, aur phir routine cases sambhalne ke liye agent ko mehdood authority de sakte hain.

Agent Factory ke alfaaz mein, Mode 1 mein automation aur augmentation zyada hain. Mode 2 agency ko munazzam banata hai. Digital FTE sirf "AI kaam kar rahi hai" nahin. Woh job definition, System of Record, permissions, rules, aur governance ke andar kaam karne wali AI hai.

Isi liye AI jitni autonomous hoti hai, 4Ds utne hi ahem ho jate hain.

Hissa 2: Chaar salahiyatein

3. Delegation: faisla karein ke kaun kya kare

Beginner ki sab se aam ghalti pehle prompt se pehle hoti hai.

Log pehle yeh faisle kiye baghair AI assistant khol kar type karna shuru kar dete hain:

• Main asal mein kya hasil karna chahta hun?
• Acha nateeja kaisa hoga?
• Kaunse hisse AI kare?
• Kaunse hisse main karun?
• Kaunse faisle AI ko kabhi na diye jayein?

Yeh Delegation ka masla hai.

Delegation ka matlab yeh faisla karna hai ke insaan aur AI ke darmiyan kaam kaise taqseem ho.

Yeh sirf "AI ko kaam dena" nahin. Yeh workflow design karna hai.

Delegation ke teen hisse hain:

1. Problem awareness: goal aur kaam ko samjhein.
2. Platform awareness: samjhein ke kaunsa AI system ya tool kaam ke liye munasib hai.
3. Task delegation: faisla karein ke har hissa kaun karega.

(Framework ka apna reference document pehle hisse ko goal and task awareness kehta hai. Idea wohi hai, aur aapko dono naam milenge.)

3.1 Problem awareness: jaanein ke kya hasil karna hai

AI se kuch maangne se pehle khud se poochein:

• Goal kya hai?
• Yeh kis ke liye hai?
• Success kaisi hogi?
• Kya ghalat ho sakta hai?
• Insani judgment kahan zaroori hai?

Farz karein aap chote business ki overdue invoices ka peecha karne ke liye AI agent chahte hain.

Beginner shayad type kare:

"Mere liye invoice-chasing agent banayein."

AI kuch bana sakti hai. Magar mushkil sawal ab bhi be-jawab hain:

• Kin customers se rabta kare?
• Invoice kitne din late ho?
• Lehja kaisa ho?
• Kis amount par insaan message approve kare?
• Customer invoice par ikhtilaf kare to kya ho?
• Agent kaunsa accounting system parh sakta hai?
• Kya agent messages bhej sakta hai, ya sirf draft kare?

Yeh prompting ke sawal nahin. Yeh business ke sawal hain.

AI aapki business policy ka faisla nahin kar sakti jab tak aap jaan boojh kar use authority na dein. Kai halat mein aapko nahin deni chahiye.

3.2 Platform awareness: sahi qisam ki AI chunein

Har AI system har kaam mein barabar acha nahin.

Aap chun sakte hain:

• mushkil multi-step masle ke liye reasoning model;
• current maloomat ke liye search-enabled assistant;
• software development ke liye coding agent;
• tools aur kai steps wale kaam ke liye agent-capable system.

Har model ka naam yaad karna zaroori nahin. Market bahut tezi se badalti hai.

Aadat yeh honi chahiye ke poochein:

"Kya yeh is kaam ke liye sahi tool hai?"

Mukhtalif systems azmayein. Results ka muqabla karein. Jo kaam kare us par notes rakhein. Which AI Employees in 2026 is badalte manzar ka current map deta hai.

3.3 Task delegation: kaam soch samajh kar baantein

Problem aur platform samajhne ke baad job ko hisson mein baantein.

Farz karein aap is jaisa course bana rahe hain:

Task	Behtareen owner	Kyun
Audience aur learning goals tay karna	Insaan	Maqsad aur judgment chahiye
Mumkin course structures sujhana	AI + insaan	AI wasee options deti hai; insaan chunta hai
Manzoor outline se sections draft karna	AI	Pehle drafts mein tez
Factual claims verify karna	Insaan	Accountability author ke paas rehti hai
Apna tajurba aur local misalein shamil karna	Insaan	AI ke paas aapka tajurba nahin
Grammar aur consistency behtar karna	AI	Mechanical review ke liye munasib
Final course approve karna	Insaan	Aapka naam aur shohrat judi hai

Yeh amli Delegation hai.

Aap "Kya AI yeh kar sakti hai?" se behtar sawal pooch rahe hain.

Aap pooch rahe hain:

"Kaunse hisse AI kare, kaunse main karun, aur kyun?"

Agent Factory mein yeh kahan le jata hai

Doosron ke liye agents banane par Delegation engineering ban jati hai.

Problem awareness specification banti hai: goal, constraints, risks, aur definition of done.

Task delegation Digital FTE ki boundary banti hai: agent kya kar sakta hai, insaan kya rakhta hai, aur kya escalate hona chahiye.

Vertical System of Record in faislon ko derpa aur inspectable banata hai.

Spec-Driven Development aur The FDE AF Model mein aage barhein.

4. Description: AI ko woh dein jo use chahiye

Course ke shuru ki colleague ko yaad karein.

Uska course outline is liye ghalat tha kyunke aapne ahem maloomat chhor di thi.

AI ka yeh masla aur gehra hai. Woh aapka zehan nahin parh sakti. Woh sirf dastiyab maloomat se kaam karti hai. Ahem baat chhor dein to woh munasib andaza laga sakti hai. Munasib andaza bhi ghalat ho sakta hai.

Description woh skill hai jismein AI ko acha kaam karne ke liye zaroori maloomat aur rehnumai di jati hai.

Description sirf "acha prompt likhne" se bahut bari hai.

Ismein teen cheezein hain:

1. Product description: aapko kya chahiye.
2. Process description: kaam kis tareeqe se ho.
3. Performance description: AI aapke saath aur khud kaise behave kare.

Seedhi yaad-dihani hai:

Kya -> Kaise -> Mere saath kaise kaam kare

4.1 Product description: nateeja wazeh karein

Product description ka sawal hai:

"Mujhe bilkul kya wapas chahiye?"

In jaisi cheezein shamil karein:

• output ki qisam;
• audience;
• format;
• length;
• tone;
• ahem topics; aur
• kya chhorna hai.

In do requests ka muqabla karein.

Ghair-wazeh:

"Is report ko summarize karein."

Zyada wazeh:

"Is quarterly financial report ko senior executives ke liye summarize karein jinke paas parhne ke liye das minute hain. Revenue trends, bare risks, aur recommended actions par tawajjoh dein. Choti bullets use karein aur ek page tak rakhein. Pichli quarter se numayan badalne wale har figure ko highlight karein. Ghair-zaroori accounting jargon se bachein."

Doosri request zyada hoshiyar nahin. Woh zyada mukammal hai.

Yeh beginner ke liye ahem lesson hai:

Aksar mukammal maloomat hoshiyar alfaaz se zyada ahem hoti hai.

4.2 Process description: tareeqa wazeh karein

Process description ka sawal hai:

"AI ko kaam kaise karna chahiye?"

Aap yeh bata sakte hain:

• kaunse steps follow karne hain;
• kaam ki tarteeb;
• kaunsa method istemaal karna hai;
• kin examples ki pairwi karni hai;
• khatam karne se pehle kaunse checks karne hain.

Misaal:

"Is code ko pehle correctness, phir security, aur aakhir mein style ke liye review karein. Jab tak yeh na jaanch lein ke code waqai chalta hai, naming issues par waqt na lagayein."

Desired product ab bhi "code review" hai. Magar ab aapne process bhi wazeh kar diya.

4.3 Performance description: rawaiya wazeh karein

Performance description ka sawal hai:

"Yeh AI kaise behave kare, aur kis ke liye?"

Aaj istemaal hone wale chote version se shuru karein, jahan "kis ke liye" ka jawab sirf aap hain.

Misaal ke taur par:

• mukhtasar ya tafseeli;
• supportive ya challenging;
• exploratory ya decisive;
• pehle sawal pooche ya munasib assumptions le;
• uncertainty flag kare ya seedha behtareen jawab de.

Ek mufeed performance description yeh ho sakti hai:

"Meri kamzor assumptions ko challenge karein. Uncertainty flag karein. Sirf adab ki wajah se mujhse agree na karein. Agar meri daleel mazboot ho to wajah batayein. Agar aapki mazboot ho to apni baat par qaim rahen aur wajah samjhayein."

Yeh AI ko adab se jawab dene wali machine ke bajaye behtar thinking partner bana deta hai.

Yeh chota version hai jo chat window mein fit hota aur sirf aapko serve karta hai. Framework isse bari cheez murad leta hai. Performance description yeh tay karna hai ke AI khud kaise behave karegi, un logon ke liye jo use karenge, unke liye bhi jo aap nahin aur aapki likhi instruction kabhi nahin dekhenge.

Dono ek hi skill ke do sizes hain. Jab aap likhte hain "sirf adab ke liye mujhse agree na karein," to aap an-dekhe behavior ka rule likh rahe hote hain. Tutoring agent ka yeh rule ke student ki koshish se pehle jawab kabhi na de, wohi jumla hai lekin stakes badal gaye: ek baar likha, hazar baar laga, aise shakhs par jisse aap kabhi nahin milenge.

Isi liye Mode 1 se Mode 2 tak jane par Description mein performance sab se zyada barhti hai. Chat mein buri performance description das minute tang karti hai. Deployed agent mein wohi product hai.

Description prompting se wasee hai

AI Prompting in 2026 examples dena, constraints batana, tasks todna, aur roles define karna jaisi amali techniques sikhata hai.

Yeh course un techniques ke peeche ka bara idea sikhata hai: Description.

Prompt ke alfaaz samajh na aa rahe hon to ruk na jayein. Apni soorat-e-haal aam zubaan mein samjhayein aur AI se use zyada wazeh instruction banane mein madad maangein.

Prompt engineering se context engineering tak

Prompt engineering poochti hai:

"Main is message ko kaise likhun?"

Context engineering bara sawal poochti hai:

"AI ki kamyabi ke liye kaunsi maloomat dastiyab honi chahiye?"

Us maloomat mein yeh shamil ho sakta hai:

• dastavez;
• misalein;
• memory;
• conversation history;
• qawaid;
• tools;
• database records;
• definitions; aur
• instructions.

Agent ke liye yeh bahut ahem hai.

Khoobsurat prompt bhi aise agent ko nahin bacha sakta jiske paas ghalat data, missing rules, bure examples, ya zaroori tools tak rasaai na ho.

Agent Factory mein yeh kahan le jata hai

Description scale ho kar architecture ban jati hai.

System prompt performance description ko derpa bana sakta hai. SKILL.md process description ko dobara istemaal ke qabil bana sakta hai. System of Record agent ke liye domain knowledge, rules, definitions, aur governance rakh sakta hai.

Us scale par achi Description, achi context engineering ban jati hai.

System of Context aur System of Record mein aage barhein.

5. Discernment: confidence ko correctness na samjhein

AI aksar confident lagti hai.

Readability ke liye yeh mufeed hai. Bharose ke liye khatarnak.

Ghalat AI jawab aam taur par warning label ke saath nahin aata. Woh polished, tafseeli, aur yaqini lag sakta hai.

Is liye kaam describe karne ke baad doosri skill chahiye:

Discernment, AI ke diye hue kaam ki quality parakhne ki salahiyat hai.

Description poochti hai:

"Kya maine job wazeh samjhayi?"

Discernment poochti hai:

"Kya AI ne waqai job achi tarah ki?"

Yahan ek mufeed istilah automation bias hai. Iska matlab automated jawab par bahut jaldi bharosa karne ka insani rujhan hai, khaas taur par jab woh confident ya professional lage.

Failure kamyabi jaisa lag sakta hai

Hallucinated jawab bilkul correct jawab jaisa lag sakta hai.

Fluent writing saboot nahin. Citation jaisa dikhne wala link saboot nahin. Confident tone saboot nahin.

Jab accuracy ahem ho, verify karein.

Discernment, Description ka aks hai. Iske bhi wohi teen hisse hain, unhi teen cheezon ki taraf:

1. Product discernment: kya nateeja acha hai?
2. Process discernment: kya AI ke saath yeh tareeqa waqai faida de raha hai?
3. Performance discernment: jab AI khud kaam kare to kya doosri taraf ke log achi tarah serve ho rahe hain?

5.1 Product discernment: kya nateeja acha hai?

Poochein:

• Kya yeh factually correct hai?
• Kya har ahem requirement follow hui?
• Kya kuch missing hai?
• Kya andarooni consistency hai?
• Kya expert ise credible samjhega?
• Kya main is par apna naam lagaunga?

Yahan aapki domain knowledge bahut qeemti ho jati hai.

Accountant ghalat accounting assumption pakarta hai. Programmer bareek bug dekhta hai. Teacher jaan leta hai ke explanation beginners ko uljha degi.

AI expert kaam ki raftaar barha sakti hai. Woh expertise ki zaroorat khatam nahin karti.

Nateeje ko parakhne ka matlab uske haq mein di gayi daleel ko parakhna bhi hai. AI kamzor wajah se sahi jawab tak pohanch sakti hai, aur ghalat assumption par khara sahi jawab der tak sahi nahin rahega.

Farz karein AI Vendor B ke muqable mein Vendor A recommend karti hai. Nateeja munasib ho sakta hai. Magar shayad usne farz kiya ho ke Vendor A mein aisa feature hai jo maujood nahin. Jab tak support durust na ho, recommendation par bharosa na karein.

AI se yeh cheezein dikhane ko kehna mufeed hai:

• pehle se mani hui baatein;
• saboot;
• faisle ke paimane;
• hisab;
• darmiyani results; aur
• mukhtalif interpretations.

Inhein review ke liye di gayi justification samjhein, model ki chhupi andarooni reasoning ka literal transcript nahin.

Misaal:

"Ek option recommend karne se pehle apni assumptions, unke haq mein evidence, aur faisle ke criteria list karein. Phir recommendation dein."

Is se jawab inspect karna asaan hota hai.

5.2 Process discernment: kya kaam ka yeh tareeqa faida de raha hai?

Kabhi jawab theek aur working relationship ghalat hota hai.

Log yeh sawal bahut kam poochte hain, kyunke output qabil-e-qabool laga aur session kamyab mehsoos hua. Aakhri message ke bajaye poore session ko dekhein:

• Kya AI mere feedback ke mutabiq dhal rahi hai, ya purane tareeqe par wapas ja rahi hai?
• Kya do baar correction ke baad bhi wohi ghalti dohra rahi hai?
• Kya woh itni agreeable ho gayi ke bekaar hai?
• Kya har turn mein wohi formatting masla theek kar raha hun?
• Kya ab main uske draft ko usse zyada edit kar raha hun jitne mein khud likh leta?

Aakhri sawal imaandari mangta hai. Bees minute ki rehnumai ek ghanta bachaye to jeet hai. Bees minute ki rehnumai sirf pandrah minute bachaye to nuqsan hai jise aapne jeet gina kyunke productive mehsoos hua.

Process na chale to teen qadmon mein barhti hui choices hain: performance description badlein, tool badlein, ya task wapas lein. Teeno fluency hain. Sirf teesra shikast lagta hai, aur aam taur par hota nahin.

5.3 Performance discernment: kya aapki ghair-maujoodgi mein AI logon ko achi tarah serve karti hai?

Teesri qisam agency ke baad hi aati hai, is liye aakhir mein book ke liye sab se ahem hai.

Performance discernment dekhti hai ke AI ka independent, user-facing behavior usse milne walon ke liye waqai ache nataij deta hai ya nahin. Yeh ek output ke correct hone se mukhtalif sawal hai.

AI tutor har jawab durust de kar bhi bura tutor ho sakta hai, agar student ke jhijakte hi solution de de aur koi kuch na seekhe. Support agent tickets jaldi close kar ke bhi bura ho sakta hai, agar woh aisi conversations band kare jinhein customers mukammal nahin samajhte.

Yeh chat window ke andar nazar nahin aata. Yeh aggregate mein dikhta hai: users agla kya karte hain, kis baat ki shikayat karte hain, aur kaunse cases baar baar khamoshi se ghalat hote hain.

Isi imaandar wajah se yeh discernment infrastructure ban jati hai. Hazar conversations aankh se nahin parhi ja sakti, is liye aap unhein dekhne wala system banate hain.

Description-Discernment loop

Description aur Discernment fitri loop banati hain:

1. Aap batate hain kya chahiye.
2. AI kuch banati hai.
3. Aap use inspect karte hain.
4. Aap batate hain kya badalna hai.
5. AI dobara koshish karti hai.

Yeh aam baat hai.

Acha AI kaam aksar iterative hota hai. Pehla jawab zyada tar draft hota hai, finish line nahin.

Feedback mein yeh seedha pattern use karein:

Masla -> Kyun ahem hai -> Direction

Kamzor feedback:

"Ghalat. Dobara karein."

Behtar feedback:

"Doosra section enterprise customers farz karta hai. Hamari audience solo founders hai, is liye mashwara bahut mehnga hai. Mehdood budget wale ek-shakhs business ke liye section dobara likhein."

Doosra version AI ko kaam ke liye mufeed maloomat deta hai.

Kabhi behtar Description kaafi nahin. Discernment dikha sakti hai ke asal Delegation decision ghalat tha. Shayad tool ghalat tha. Shayad task ka woh hissa AI ke paas hona hi nahin chahiye tha. Shayad insani expert chahiye.

Yeh bhi fluency hai.

Agent Factory mein yeh kahan le jata hai

AI systems banate waqt Discernment evaluation engineering ban jati hai.

Aapka manual sawal, "Kya yeh kaafi acha hai?", eval suites, production checks, monitoring, sampling, aur release gates mein badal jata hai.

Aap aisi judgment automate nahin kar sakte jo khud karna kabhi seekhi hi na ho.

Trusting the Checker aur Eval-Driven Development mein aage barhein.

6. Diligence: AI ko zimmedari se use karein aur nateeja apnayein

Pehle teen Ds aapko behtar results dilati hain.

Diligence mukhtalif sawal poochti hai:

"Kya mujhe AI ko is tarah use karna bhi chahiye?"

Sochiye lecturer AI se students ke end-of-term feedback ka draft banwata hai.

Writing behtareen hai. Magar lecturer ne students ke naam, grades, aur disciplinary notes consumer AI service mein paste kiye jise university ne kabhi approve nahin kiya. Students ko bataya bhi nahin gaya ke comments mein AI ka hissa tha jo unke academic record ka hissa ban sakte hain.

Output acha ho sakta hai.

AI ka istemaal phir bhi ghair-zimmedarana hai.

Diligence ka matlab AI ke istemaal aur uske output ke nataij ki zimmedari lena hai.

Diligence ke teen hisse hain:

1. Creation diligence: banane se pehle aur dauran zimmedarana choices karein.
2. Transparency diligence: jab logon ko jaanna chahiye to AI ke kirdar ke baare mein imaandar rahen.
3. Deployment diligence: kaam istemaal ya release hone se pehle verify karein aur zimmedari lein.

6.1 Creation diligence: tools aur data zimmedari se chunein

AI system ke saath maloomat share karne se pehle poochein:

• Kya ismein personal data hai?
• Kya ismein confidential company information hai?
• Kya mujhe yeh maloomat is tool mein daalne ki ijazat hai?
• Data tak kaun pohanch sakta ya use rakh sakta hai?
• Kya meri organization ne yeh service approve ki hai?
• Kya qanooni, contractual, ya professional pabandiyan hain?

Sab se asaan raasta hamesha zimmedarana nahin hota.

Asal customer database ko khule hue kisi bhi AI tool mein copy karna paanch minute bacha kar privacy ya compliance ka sanjeeda masla paida kar sakta hai.

6.2 Transparency diligence: AI ke kirdar par imaandar rahen

Har AI-assisted task ko public announcement nahin chahiye.

Magar jab AI doosre logon par numayan asar dale, disclosure ahem ho sakta hai.

Misalein:

• taleemi kaam;
• bharti ke faisle;
• customers se rabta;
• medical ya financial advice;
• professional reports; aur
• insaan ka asal kaam bana kar pesh kiya gaya content.

Theek rule context, organization, qanoon, aur professional standard par munhasir hai.

Ek mufeed usool hai:

AI-assisted nateeja jitna doosre logon ko mutasir kare, transparency ki wajah utni mazboot hoti hai.

Transparency ka matlab workflow ki har tafseel batana nahin. Matlab yeh hai ke jahan AI ka kirdar ahem ho, wahan logon ko iske baare mein gumrah na karein.

6.3 Deployment diligence: haath se nikalne se pehle verify karein

AI-assisted kaam publish, send, execute, ya kisi faisle mein istemaal hone se pehle use jaanchein.

Task ke mutabiq iska matlab ho sakta hai:

• facts verify karein;
• tasdeeq karein ke sources waqai maujood hain;
• calculations check karein;
• bias ya na-insaf nataij review karein;
• permissions aur rights confirm karein;
• organization policy follow karein;
• high-impact actions ke liye insani approval lein.

Aakhri powerful sawal hai:

"Kya main itminan se is par apna naam lagaunga?"

Jawab nahin ho to kaam taiyar nahin.

Diligence ka sab se ahem lesson seedha hai:

AI kaam automate kar sakti hai. Accountability automate nahin kar sakti.

AI-assisted system nuqsan-deh faisla kare to use chalane wali organization phir bhi zimmedar hai. Coding assistant vulnerability shamil kare aur engineer ship kare to engineer aur organization nateeje ke malik hain.

Agent Factory mein yeh kahan le jata hai

Diligence ki wajah se Agent Factory governance-first hai.

System scale par:

• Creation diligence data rules, access control, aur approved-tool policy banti hai.
• Transparency diligence disclosure aur user experience design banti hai.
• Deployment diligence evaluation gates, audit logs, monitoring, aur human review banti hai.

Mode 2 mein aap sirf khud responsible AI use nahin karte. Aap zimmedari ko aise product mein banate hain jo doosre log istemaal karenge.

System of Record aur Designing the Vertical SoR mein aage barhein.

Hissa 3: 4Ds ko saath istemaal karein

7. 4Ds ek amali operating loop ke taur par

Aapne chaaron salahiyatein alag seekhin. Asal kaam mein yeh mil jati hain.

Framework ke authors Description-Discernment loop wazeh taur par sikhate hain. Agent Factory ke liye yeh book idea ko barhati aur chaaron salahiyaton ko ek amali operating loop banati hai:

Delegate -> Describe -> Discern -> Zimmedari nibhayein -> Zaroorat par dohrayein

Har hissa yeh deta hai:

• Delegation tay karti hai ke AI kaam mein aaye ya nahin aur kis cheez ki malik ho.
• Description AI ko goal, context, process, aur behavior deti hai.
• Discernment nateeja check kar ke agla round behtar karti hai.
• Diligence poore process ko zimmedari se gher leti hai.

Ab ek asal misaal par loop dekhein.

Misaal: bookkeeping Digital FTE

Ayesha Lahore mein Forward Deployed Engineer hai. Woh Karachi ki choti accounting practice ko bookkeeping Digital FTE banane mein madad de rahi hai.

Pehli job jise woh automate karna chahte hain monthly bank reconciliation hai.

Qadam 1: Delegation

Ayesha AI se "reconciliation agent banayein" keh kar shuru nahin karti.

Pehle accounting partners ke saath job map karti hai.

Woh tay karte hain:

• agent bank transactions ko ledger entries se match kar sakta hai;
• agent unmatched items flag kar sakta hai;
• agent reconciliation report draft kar sakta hai;
• har journal adjustment insaan approve karega;
• har write-off decision insaan ke paas rahega;
• client ki tax position ko mutasir karne wali har cheez accountant ke paas rahegi;
• high-value unmatched items naamzad shakhs ko escalate honge.

Ab boundary wazeh hai.

Qadam 2: Description

Phir Ayesha system ko zaroori maloomat deti hai:

• firm ka chart of accounts;
• matching ke qawaid;
• pichli reconciliations ke examples;
• partners ka pehle se istemaal hota report format;
• escalation ke qawaid;
• duplicate payments aur stale cheques ki definitions;
• rule ke agent khud journal entry kabhi post nahin karega;
• rule ke agent client se seedha rabta kabhi nahin karega.

Yeh system scale par Description hai.

Qadam 3: Discernment

Ayesha impressive demo dekh kar agent ko kamyab farz nahin karti.

Woh use pichli un reconciliations par test karti hai jin par firm ko bharosa hai.

Team jaanchti hai:

• kitne matches correct hain;
• kitne incorrect matches nikal jate hain;
• kya sahi cases escalate hote hain;
• kya agent bahut zyada escalate karta hai;
• kya performance waqt ke saath badalti hai.

Accountant sirf failures nahin, zahiran kamyab matches bhi review karta hai. Yeh ahem hai kyunke system khamoshi se fail ho kar mehfooz nazar aa sakta hai.

Qadam 4: Diligence

Client financial data approved infrastructure ke andar rehta hai.

Agent actions log hote hain.

Jahan disclosure zaroori ya munasib ho, clients ko bataya jata hai ke reconciliation AI-assisted hai.

Insani partner ab bhi reconciliation sign karta hai.

Final nateeje ki accountability partner ke paas rehti hai.

Yeh amli 4D loop hai.

Zaati skill system property ban gayi hai.

Salahiyat	Chat mein	Agent Factory mein
Delegation	Faisla ke AI se kya karwana hai	Digital FTE aur human/AI boundary scope karna
Description	Instructions aur context dena	System prompts, skills, context engineering, Systems of Record
Discernment	Jawab review karna	Evals, monitoring, sampling, trusting the checker
Diligence	Data bachana aur nateeja apnana	Governance, permissions, audit, disclosure, human review

Agent Factory AI fluency ki jagah nahin leti.

Woh use industrialize karti hai.

Iska Four Survival Skills aur 10-80-10 Rule se talluq

4Ds is book ke ideas ka personal-scale version hain.

10-80-10 Rule ko 4Ds ke zariye samajhna asaan hai:

• Pehla 10%: direction set karein. Yahan Delegation aur Description mazboot hain. Faisla karein kya qabil-e-kaam hai aur goal wazeh karein.
• Darmiyani 80%: AI ko orchestrate karein. AI kaam banati aur aap rehnumai karte hain to Description aur Discernment musalsal dohrati hain.
• Aakhri 10%: sach parakhein. Ahem cheez ship hone se pehle Discernment intehai zaroori hai.
• Poore 100% mein: zimmedari se kaam karein. Diligence aakhri checkbox nahin. Woh poore workflow ko gher leti hai.

Is liye chat window mein 4Ds ki practice karke aap wohi insani skills rehearse karte hain jo baad mein AI systems banane aur govern karne mein chahiye hongi.

8. Beginners ki chaar aam ghaltiyan

Zyada tar mayoos-kun AI tajurbe in failures mein se ek tak pohanchte hain.

Ghalti 1: Masla wazeh karne se pehle prompting

Aap success ki shakal tay karne se pehle type karna shuru kar dete hain.

Jis skill ki kami hai: Delegation

Hal: pehle goal, audience, constraints, aur human/AI split wazeh karein.

Ghalti 2: Pehle jawab ko final jawab samajhna

Aap farz karte hain ke kamzor pehla jawab sabit karta hai ke AI bekaar hai.

Jis skill ki kami hai: Description + Discernment loop

Hal: nateeja inspect karein, khaas feedback dein, aur iterate karein.

Ghalti 3: Polished jawab ko professional lagne ki wajah se maan lena

Aap fluency ko accuracy samajh lete hain.

Jis skill ki kami hai: Discernment

Hal: ahem facts, assumptions, calculations, aur sources verify karein.

Ghalti 4: Privacy ya accountability par tab sochna jab kuch ghalat ho jaye

Aap task mukammal karne par tawajjoh dete aur AI ke istemaal ko nazar-andaz karte hain.

Jis skill ki kami hai: Diligence

Hal: deployment se pehle data, disclosure, approval, aur accountability ke rules tay karein.

Roz istemaal hone wali beginner checklist

AI ko meaningful kaam dene se pehle yeh chota check karein:

Stage	Khud se poochein
Delegate	Goal kya hai? AI kya kare? Mere paas kya rahe?
Describe	AI ko kaunsa output, context, method, aur behavior chahiye?
Discern	Main kaise jaanunga ke jawab correct, mukammal, aur mufeed hai?
Diligence karein	Kya data mehfooz hai? Kya AI ka kirdar batana hai? Nateeja kaun approve aur own karega?

Har chote task ke liye ise paperwork banana zaroori nahin.

Maqsad chaar sawalon ko khudkaar aadat banana hai.

Practice se pehle mukhtasar khulasa

AI fluency prompts yaad karne ki salahiyat nahin. Woh AI ke saath effective, efficient, ethical, aur safe tareeqe se kaam karne ki salahiyat hai.

Aap AI ke saath teen modes mein kaam kar sakte hain:

• Automation: AI defined task karti hai.
• Augmentation: aap aur AI mil kar sochte hain.
• Agency: AI aapke set kiye goal ki taraf khud kaam karti hai, aksar un logon ke liye jo aap nahin.

Teeno modes mein chaar salahiyatein wohi rehti hain:

• Delegation: human/AI split tay karein.
• Description: AI ko zaroori cheezein dein.
• Discernment: wapas aane wala kaam parakhein.
• Diligence: AI zimmedari se use karein aur nateeja apnayein.

Agar course ka sirf ek jumla yaad rakhein, to yeh:

Faisla karein AI ko kya karna chahiye. Kaam wazeh batayein. Wapas aane wala kaam jaanchein. Agle nateeje ki zimmedari lein.

Ab yeh azmayein: chhe prompts

AI fluency ke baare mein parhna kaafi nahin. Use karein.

AI assistant kholein aur neeche ki exercises azmayein. Chhe aik hi sitting mein karna zaroori nahin.

1. Asal task ke liye 4D plan banayein

Woh cheez chunein jo is haftay waqai karni hai.

I need to do this: [describe the task].

Before we start, walk me through the four Ds of AI fluency for it:
delegation, description, discernment, diligence. Ask me one question
at a time, skip any that obviously does not apply, and give me the
plan as a short table at the end.

Kya dekhein: plan zyada tar AI ke nahin, aapke jawabon se bana hai. Yahi maqsad hai. Delegation aur Diligence ke faisle sirf aap kar sakte hain. AI yahan sawal yaad dilane mein mufeed hai, faisla karne mein nahin.

2. Us topic par Discernment use karein jo aap jaante hain

Aisa subject chunein jahan aapko asal tajurba hai.

Let's discuss [topic I know well].

Talk to me like a knowledgeable colleague, not a lecturer.
As we go, I will watch for three things:
- where you improve my thinking,
- where I need to correct you,
- where my own experience makes me reject your suggestion.

Start by asking which part of the topic I want to discuss.

Kya dekhein: domain jaante hon to discernment kitni sasti hai. Aapne ghalat claim baghair mehnat, taqreeban baghair faisle ke pakar liya. Yeh aapki expertise ka kaam hai, aur exercise 3 mein bilkul yehi cheez aapke paas nahin hogi.

3. Non-expert hona mehsoos karein

Ab aisa topic chunein jiske baare mein taqreeban kuch nahin jaante.

Teach me the basics of [topic I know little about].
Explain it for a complete beginner and use concrete examples.

At the end, identify the claims in your explanation that I should
verify with a reliable source, and explain why they deserve checking.

Kya dekhein: wohi quality ka output kitna mukhtalif lagta hai. Kuch ghalat mehsoos nahin hua, kyunke jaanchne ke liye aapke paas kuch tha hi nahin. Is ehsas ko yaad rakhein: aapke banaye agent ka har user hamesha aisa hi mehsoos karega, aur isi liye Digital FTE ko trust ke bajaye evals chahiye.

4. Performance description likhein

Serious working session ke shuru mein yeh use karein:

During this conversation:
- challenge weak assumptions,
- flag uncertainty on factual claims,
- do not agree with me just to be polite,
- ask a clarifying question when an ambiguity would materially change the answer,
- change your recommendation when new evidence supports a change,
- explain why when you disagree with me.

Kya dekhein: farq kitne kam turns mein wazeh hota hai, aur fresh chat khol kar yeh dobara set na karein to asar kitni jaldi mit jata hai. Isi liye deployed agent performance description kisi ki memory ke bajaye system prompt mein rakhta hai.

5. Recommendation maan-ne se pehle justification inspect karein

AI ko apna asal faisla dein, phir yeh shamil karein:

Before making a recommendation, list:
1. the important assumptions,
2. the evidence supporting them,
3. the criteria you are using to compare the options,
4. the major uncertainties.

Then make the recommendation.
I want a justification I can review, not just a conclusion.

Kya dekhein: kya list mein koi assumption aisi hai jise likhe baghair aap khamoshi se maan lete. Wohi check karne layak hai, aur jab AI sirf nateeja de to woh nazar nahin aati.

6. Chota project poore 4D loop se guzarein

Ek ghante mein mukammal hone wala project chunein: study plan, tutorial, presentation, proposal, ya chota coding task.

I want to complete this project using the 4D AI Fluency framework:
[describe the project].

First, help me decide the human/AI division of work.
Then, before each AI-owned task, ask what product, process,
and performance I want.

After each important output, stop so I can evaluate it.
At the end, run a diligence check covering facts, sensitive data,
disclosure, approvals, and anything I should verify before using the work.

Khatam kar ke khud se poochein:

Kaunsi D ne mujhse sab se zyada mehnat mangi?

Shayad wohi competency hai jiski sab se zyada practice chahiye.

Tez self-check

Peeche dekhe baghair yaad se jawab dein. Kisi sawal par scroll karna pare to woh section abhi zehan mein nahin baitha.

1. AI fluency ko kaunsi chaar qualities define karti hain?
2. Automation, augmentation, aur agency mein kya farq hai?
3. Delegation ke teen hisse kya hain?
4. Description ke teen hisse kya hain?
5. Confident AI jawab ko verify karna phir bhi kyun zaroori hai?
6. Discernment ke teen hisse kya hain?
7. Diligence ke teen hisse kya hain?
8. AI-assisted kaam ship karne se pehle kaunsa sawal pooch sakte hain?
9. Ek jumle mein 4D loop kya hai?
10. Agent Factory mein Discernment engineering practice kaise banti hai?

Jawabat

1. Effective, efficient, ethical, aur safe.
2. Automation defined task karti hai, augmentation thinking partner ke taur par aapke saath kaam karti hai, aur agency zyada azadi se steps chun kar goal ki taraf kaam karti hai.
3. Problem awareness, platform awareness, aur task delegation.
4. Product description, process description, aur performance description.
5. Kyunke AI plausible output banati hai, aur plausible correct ke barabar nahin. Fluent alfaaz facts, assumptions, ya reasoning verify nahin karte.
6. Product discernment, process discernment, aur performance discernment.
7. Creation diligence, transparency diligence, aur deployment diligence.
8. "Kya main itminan se is par apna naam lagaunga?"
9. Faisla karein AI ko kya karna hai, kaam describe karein, wapas aane wale kaam ko parakhein, aur poore process ki zimmedari lein.
10. Woh evals, monitoring, sampling, review gates, aur AI system ki performance jaanchne ke doosre tareeqe banti hai.

Is course ki nayi istilahat

Machine ki technical vocabulary What AI Actually Is aur book ki Glossary mein hai. Is course ki ahem istilahat yeh hain.

AI fluency. AI ke saath effective, efficient, ethical, aur safe tareeqe se kaam karne ki salahiyat.

4Ds. Delegation, Description, Discernment, aur Diligence.

Automation. AI khaas instructions se defined task karti hai.

Augmentation. Insaan aur AI thinking partners ke taur par mil kar kaam karte hain.

Agency. AI kisi shakhs ki taraf se goal ki taraf kaam karti aur kai steps khud chunti hai.

Delegation. Faisla ke kaunsa kaam hona hai, AI kya kare, aur insaan kya apne paas rakhein.

Problem awareness. AI ko shamil karne se pehle goal, kaam, risks, aur success ka matlab samajhna.

Platform awareness. Samajhna ke kaunsa AI system ya tool task ke liye munasib hai.

Task delegation. Kaam ke hisse soch samajh kar insaan ya AI ko dena.

Description. AI ko acha kaam karne ke liye zaroori maloomat aur rehnumai dena.

Product description. Desired output define karna.

Process description. AI ka kaam karne ka tareeqa define karna.

Performance description. Define karna ke AI khud un logon ke liye kaise behave kare jo use karenge. Chat scale par yeh aapki working-style preferences hain; system scale par deployed agent ka standing behavior.

Discernment. Parakhna ke AI ka output, justification, aur behavior kaafi ache hain ya nahin.

Product discernment. Nateeje ko khud parakhna.

Process discernment. Ek output ke acha hone ke bajaye parakhna ke AI ke saath kaam ka tareeqa waqai faida de raha hai ya nahin.

Performance discernment. Parakhna ke AI ka independent, user-facing behavior serve hone wale logon ke liye ache nataij deta hai ya nahin.

Diligence. AI ke istemaal aur uske output ke nataij ki zimmedari lena.

Creation diligence. Banane se pehle aur dauran tools, data, aur AI use zimmedari se chunna.

Transparency diligence. Jab mutasir logon ke liye AI ka kirdar ahem ho to uske baare mein imaandar hona.

Deployment diligence. AI-assisted kaam ke istemaal, publication, sending, ya execution se pehle verify karna aur zimmedari lena.

Context engineering. AI system ke liye zaroori maloomati mahaul design karna: instructions, documents, tools, memory, policies, examples, aur doosra relevant context.

Automation bias. Automated output par bahut asani se bharosa karne ka insani rujhan.

Hallucination. Confident ya plausible AI output jismein banayi hui ya ghalat maloomat ho.

Aage kahan jana hai

Ab aapke paas woh insani framework hai jis par baqi Foundations banti hain.

Agla course AI Prompting in 2026, Description ko roz ki amali aadaton mein badalta hai.

How to Think in the AI Era, Discernment aur critical thinking ko mazboot karta hai.

Build path mein aage:

• Spec-Driven Development, Delegation aur Description ko engineering method banata hai.
• Trusting the Checker aur Eval-Driven Development, Discernment ko evaluation infrastructure banate hain.
• System of Record aur Designing the Vertical SoR, Diligence, context, aur governance ko system architecture banate hain.

Agar credentials jama kar rahe hain to un exams ke liye Certifications dekhein jin ki taiyari yeh material karwata hai.

Sources aur license note

AI Fluency Framework Rick Dakan, Ringling College of Art and Design mein Professor of Creative Writing aur AI Coordinator, aur Joseph Feller, Cork University Business School, University College Cork mein Professor of Information Systems and Digital Transformation, ne banaya. Unka course AI Fluency: Framework and Foundations Anthropic ke saath bana aur CC BY-NC-SA 4.0 ke tehat release hua. Unka reference document, Framework for AI Fluency: Practical Overview Document, CC BY-NC-ND 4.0 ke tehat release hua.

Yeh crash course framework ki is book ke apne alfaaz aur misalon mein mustaqil wazahat hai. Yeh unke materials ki copy ya adaptation nahin. Yahan istemaal hui har competency, sub-competency, aur modality definition ko authors ke Practical Overview Document aur framework terminology sheet ke saath check kiya gaya. Agent Factory mappings, har competency ka chat se factory tak scale, aur four-D operating-loop framing is book ki extensions hain. Un extensions mein koi ghalti hamari hai, unki nahin.

Asal bhi parhein. Woh muft hai, aur framework ke peeche research ko kisi bhi summary se behtar samjhata hai.

• The AI Fluency Framework, authors ki apni site, papers, presentations, aur open educational resources ke saath.
• AI Fluency: Framework and Foundations, Claude Academy par muft 12-lesson asal course, taqreeban 3 se 4 ghante. anthropic.skilljar.com par bhi dastiyab.
• Framework for AI Fluency: Practical Overview Document, authors ka living reference document aur sub-competencies ka sab se durust bayan.
• Companion courses Teaching AI Fluency, AI Fluency for Educators, aur AI Fluency for Students, Claude Academy aur OpenCourses.ie par muft.
• The AI Fluency Framework terminology sheet (PDF), copyright 2025 Rick Dakan, Joseph Feller, aur Anthropic, CC BY-NC-SA 4.0 ke tehat release hua.

---

Flashcards Se Parhai Mein Madad

---

Apni samajh azmayein

Course ke oopar ke chaar sawal parhne mein asaan aur asal mein chalane mein mushkil the. Yeh scenarios aapko doosron ke kaam mein wahan daalte hain jahan faisla pehle se muntazir hai. Alfaaz ke bajaye reasoning se jawab dein, aur note karein ke har sawal asal mein kaunsi D test karta hai.

Checking access...