• Foundation
  • About 15 min
  • 5 practice questions

Jobs, skills and staying safe from AI-driven fraud

Explain how AI changes and creates jobs, which skills to build, and how to protect yourself and your family from deepfakes, voice clones and AI-written scams, including reporting on 1930 or cybercrime.gov.in.

If something looks wrong, please tell us.

Every big technology changes work. Computers removed many typing-pool jobs but created jobs in software, data entry and IT support. AI is doing the same, faster. At the same time, the tools that make realistic voices and videos have handed criminals new tricks. This last teaching lesson covers both sides: how to prepare for work alongside AI, and how to stay safe from fraud that uses it.

How AI changes work

Three ways AI affects jobs
Kind of changeWhat it meansExamples
Tasks automatedRepetitive, rule-based tasks are done by softwareCopying data from forms, sorting routine emails, basic transcription
Jobs changedPeople keep the job but use AI tools within itDoctors reading AI-flagged scans, teachers using adaptive apps, clerks using document-reading tools
Jobs createdNew work appears around building and checking AIData annotators who label examples, data analysts, AI engineers, people who test AI for errors and bias, drone operators

Notice that AI usually takes over tasks, not whole jobs. Work that needs judgement, care, trust, negotiation or hands-on skill in changing surroundings is much harder to automate.

Which tasks AI takes over

  • Easier to automate
    • Repetitive, rule-based tasks
    • Copying data from forms
    • Sorting routine emails
    • Basic transcription
  • Harder to automate
    • Judgement
    • Care and trust
    • Negotiation
    • Hands-on skill in changing surroundings

Skills worth building

  • Data literacy: reading tables and charts, asking where data came from and whether it is complete.
  • Basic programming: writing simple programs; Python is widely used in AI and is a common starting language.
  • Basic mathematics and statistics: averages, percentages and probability help you understand what an AI's 'confidence' means.
  • Critical thinking: checking AI output instead of trusting it, and spotting when an answer sounds confident but is wrong.
  • Communication and domain knowledge: knowing farming, health, law or accounts well, so you can use AI tools wisely in that field.
  • Learning to learn: tools change quickly, so the habit of picking up new ones matters more than any one tool.

AI-driven fraud: what to watch for

Common AI-driven frauds
TrickHow it worksWarning signs
Voice cloneA few seconds of someone's recorded voice are used to make a fake call in that voiceAn urgent request for money, a new number, pressure not to tell anyone
Deepfake videoAI makes a realistic video of a known person saying things they never said, often promoting an 'investment'Promises of guaranteed or doubled returns, links from unknown accounts, odd lip movement
AI-written messagesFluent, personalised fake messages from a 'bank', 'courier' or 'officer'Requests for OTP, PIN, card details or app downloads; threats of account closure or arrest

Deepfakes keep improving, so do not rely only on spotting visual flaws. The stronger defence is to verify through a channel you already trust.

Worked example 1: the panicked call

Sunita gets a call in her son's voice: 'Ma, I've had an accident, send ₹50,000 to this UPI ID now, don't tell Papa.' Working: urgency, secrecy and a new payment ID are classic warning signs. Step 1, she stays calm and does not pay. Step 2, she hangs up and calls her son on the number saved in her phone. Step 3, if the family has agreed a code word, she asks for it. Her son answers from college, safe. The call was a voice clone. Step 4, she reports the number. Had she already paid, she should call 1930 at once, because quick reporting improves the chance of stopping the money.

Sunita's four steps

  1. Stay calm, do not pay
  2. Call back on the saved number
  3. Ask the family code word
  4. Report the number Had she paid, call 1930 at once.

Worked example 2: a celebrity 'investment' video

Ravi sees a video on social media in which a famous business leader seems to promise that a new app will double any money within a month. Working: guaranteed high returns are a red flag on their own; the link comes from an unknown page; and the claim does not appear on the person's official channels or in any reliable news source. Conclusion: very likely a deepfake scam. Ravi does not click, and reports the video on the platform and at cybercrime.gov.in.

Pause, verify, report. Never share an OTP or PIN. Report cyber fraud on the helpline 1930 or at cybercrime.gov.in, as quickly as possible.

Key takeaways: AI automates tasks, changes most jobs and creates new ones in data and AI work. Data literacy, basic programming, critical thinking and domain knowledge are the skills to build. Voice clones, deepfakes and AI-written messages make scams convincing, so verify through a trusted channel, use a family code word, never share OTPs, and report quickly on 1930 or cybercrime.gov.in.

Check your understanding

  1. 1. Which kind of work is AI most likely to automate?

    1. Comforting a patient's family
    2. Repetitive copying of data from forms
    3. Negotiating a peace agreement
    4. Repairing a pipe in a crowded house
    Show answer

    Answer: B. AI and software handle repetitive, rule-based tasks such as copying data from forms best. Work needing care, judgement, negotiation or hands-on skill in changing surroundings is much harder to automate.

  2. 2. A data annotator's job is to:

    1. Label examples, such as tagging photos, so AI systems can learn from them
    2. Repair mobile phones
    3. Sell insurance
    4. Drive delivery vans
    Show answer

    Answer: A. AI learns from labelled examples, and data annotators create those labels, for instance marking which photos show a diseased leaf. It is one of the jobs AI has created. Phone repair, insurance sales and driving are unrelated.

  3. 3. You get a call in a relative's voice urgently asking for money on a new UPI ID. What is the best first action?

    1. Pay quickly, since the voice sounds right
    2. Share your OTP to speed up the transfer
    3. Send half the amount to be safe
    4. Hang up and call the relative back on the number you already have saved
    Show answer

    Answer: D. Voices can be cloned, so a familiar voice proves nothing. Verify through a channel you already trust: call back on the saved number or ask a family code word. Paying, sharing an OTP or sending part of the money all help the fraudster.

  4. 4. What is the national helpline number for reporting cyber fraud in India?

    1. 100
    2. 112
    3. 1930
    4. 1098
    Show answer

    Answer: C. The cyber fraud helpline is 1930, and complaints can also be filed at cybercrime.gov.in. 112 is the emergency response number, 100 is the police number, and 1098 is Childline.

  5. 5. Which skill helps most in checking whether an AI tool's confident answer is actually right?

    1. Typing speed
    2. Critical thinking
    3. Handwriting
    4. Memorising phone numbers
    Show answer

    Answer: B. Critical thinking means questioning output, checking sources and noticing when something sounds confident but may be wrong. Typing speed, handwriting and memorising numbers do not help judge an answer's correctness.

Privacy · Terms · Refunds · Delivery