Fair game: the ethics of telco fraud
summary
An ethics ops consultant embeds with a small telco fraud team using simple deterministic rules plus a third‑party ML payment gateway, chasing international revenue‑sharing and promotion abuse. Concrete takeaways: alerts flow into Slack with batch delays that cause false alarms, a 30‑in‑24h threshold can mean different units, and a high SMS count once turned out to be manual group chats not bots. Priorities are deconstruct your proxy, explicit data design, and failure‑first design with in‑product feedback and recourse.
AI-written summary · about
chapters (16)
- 0:00Intro
- 1:59Fair game The ethics of telco fraud
- 2:35Setting the scene
- 5:02A balancing act
- 5:34Developing a healthy cynicism about data sources
- 5:59Data (interpretation) pride comes before a fall
- 10:48Data misinterpretation is easy, and likely under pressure
- 12:38Pointers for data interpretation
- 15:02FINDING #3 Assume good faith
- 17:56Keep the moralising out of it
- 18:37Tip Describe behaviours, not people
- 20:29"Assuming perfect knowledge of the objective decouples the machine from the human: what the human does no longer matters, because the machine knows the goal and pursues it." -Stuart Russell, Human Compatible
- 24:49Think the interaction through • Plan for customer support
- 25:44People don't like it when you make bad assumptions about them!
- 26:501. Deconstruct your proxy 2. Explicit data design 3. Assume good faith 4. Failure-first design
- 28:06Sweet Summer Child Score
related talks
29mGuardians of the Gateway: Battling SMS Fraud in Real Time
20mFraud fighting: leverage real-time network intelligence
32mHow Top Operators Fight Fraud in Crypto & Gaming
28mUnderstanding IRSF Fraud: Protecting Against...
23mTo Catch A Fraudster: Building Products for the Dynamic World of Online Fraud
42mThe Old is New Again: An Intro to SMS Security and Why It’s Vital
45mModern fraud trends in RTB
45mAnti-Fraud in Reverse and the Use of ML Methods in It