The harms land on people who did not take this course
Almost everything in these nine modules affects people who never chose the system, cannot see it, and do not know what to ask. Which means the most useful thing most readers will ever do with this material is help somebody else: a parent who got the phone call, a teenager whose images are circulating, a neighbour whose benefit stopped, a friend whose account was deactivated.
The first ten minutes, whatever the harm
Three things, in this order.
Believe them. Help them pause any demand from the sender. If money has already been sent or someone is in immediate danger, get help straight away.
Preserve evidence before anything else. Screenshots including the full screen with date and time, URLs, account names, reference numbers, message headers, the exact wording. Save it outside the account in question. People delete things to make them stop existing, and it destroys their case.
Do not shame them. Fraud victims often do not report because they feel foolish, and the successful ones are engineered by professionals against ordinary people. A person who feels judged stops telling you things, and you need what they have not said yet.
Then, by type
Money has been sent. Contact the bank or payment provider immediately. In India, call 1930 for financial cyber fraud and use the National Cyber Crime Reporting Portal at cybercrime.gov.in. In England, Wales and Northern Ireland, use Report Fraud. In Scotland, use Police Scotland. Keep transaction details and messages privately for the bank and reporting service. Be alert for a follow-up message offering to recover the money.
A voice or video was used to impersonate someone. Verify by an independent channel before anything else. Then warn the person being impersonated, who may be being used against others too.
Intimate images, real or synthetic. Do not have them send you the images. Preserve URLs and account names only. Use the platform’s specific non-consensual intimate image reporting route, which is faster than a general report. Hash-based services such as StopNCII let a person register an image so participating platforms can block matching uploads without the image being sent anywhere. Report to police even where the law is unclear, because the record matters later. If a minor is involved, this is a child protection matter and goes to the police and the relevant national reporting line, not to the school’s discretion alone.
A decision was made about them. The letter from the fifth module. Find the deadline first. Ask whether the decision was solely automated, request human review by someone uninvolved, ask what data was used and what would have had to be different, and dispute specific facts with evidence. Keep everything in writing.
Their work or account is gone. Data access request, request for human review, screenshots of everything before access is lost, and find whether a collective exists: worker groups have obtained far more than individuals have.
They are relying on a chatbot in distress. Do not remove it and offer nothing. Ask what it gives them: usually availability, no judgement, or the ability to say something out loud. Then find a human route that supplies the same thing, and give the local crisis line. If there is any indication of immediate risk, that is a call to emergency services or a crisis service, now, not a conversation about technology.
What to say when you do not know
Three honest sentences cover most of it:
”I do not know, and here is who would.” Name the regulator, the union, the legal aid clinic, the cybercrime portal, the data protection authority.
Keep a dated record and the relevant messages, links or transaction details. Make any urgent call first, or collect the evidence while someone else makes the call.
”You are not the first person this happened to.” True in every case, and it is the sentence that gets people to act rather than hide.
The end of the course
Eight modules of mechanism and one of practice, and the through-line is a single move: ask who chose, what they measured, who bears the cost, and what happens when it is wrong.
That question works on a hiring model, a benefits system, a chatbot, a face-recognition alert, a training corpus and a data centre. It does not require you to be a technologist. It requires you to refuse the framing in which a decision made by people, for reasons, with consequences that fall on somebody, gets described as something a machine did.
Check your understanding
Someone tells you a relative just transferred money after a distressed call in their daughter’s voice. What is the correct order of first actions?
Pick the one you would defend. Your answer stays private.
Your place and first choices stay on this browser. Classroom credit is recorded separately.
Stuck on something?Ask Rukmini, your AI tutor