Scribd
Upload
42 views2 pages

Chapter 4 Part 2 - Answers

Uploaded by

Robert Oo
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
42 views2 pages

Chapter 4 Part 2 - Answers

Uploaded by

Robert Oo
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 2

Tutorial 4 Part 2 Answer

Chapter 4 Part 2 (Artificial Intelligence)


Question 1. Define artificial intelligence (AI). What are the long-term goals of AI?
Answer
Definition
Artificial intelligence (AI) represents a computational approach to developing intelligent
software and systems that can solve complex problems by adopting human intelligence and
working in the ways of human mind.
Long-term goals of AI
The long-term goal of AI is to achieve artificial general intelligence (AGI). AGI is a strong
form of AI and can be aptly described as the idea that “the appropriately programmed
computer with the right inputs and outputs would thereby have a mind in exactly the same
sense human beings have minds.”
Question 2. Discuss ‘fraud detection’ as present and future applications of AI in financial
services.
Answer
The frequency of online financial fraud attempts has been increasing over time. Fraud
detection is tricky, as fraud means change from time to time. One major challenge is to detect
“Black Swan”–like fraud events, which are events that rarely happen but will bring about
catastrophic losses if they happened. Fraud detection systems in the past were rule-based.
The main defect of a rule-based system is that it cannot detect unidentified fraud means. As
there is little data on such Black Swan–like fraud cases due to their low frequency of
occurrences, detection using the old rule-based system is nearly impossible.
Modern advanced fraud systems adopt sophisticated machine learning models to detect the
potential fraud, which implies that fraud can be flagged before it happens. Rather than
general machine learning black boxes, a fraud system is called “white boxing,” which
consists of a scoring system based on local linear approximation, a text mining system, graph
generation, and a visualization system. It can thus detect emerging patterns by learning new
data and generate a probability of potential fraud.
Question 3. Discuss ‘loan / insurance underwriting’ as present and future applications of AI
in financial services.
Answer
Many AI-enabled credit assessment systems have emerged and small loan underwriters tend
to employ such an end-to-end AI credit system to assess borrower’s credit using large
amounts of data from social and ecommerce media, which has not been traditionally used for
credit evaluation. The customers’ behavioral traits from social media such as the contents of
text messages, the patterns of call history, geolocation information, and mobile transaction,
are valuable information. Such information provides insight into customers’ identity,
incomes, and expenses. The system adopts algorithms and machine learning to find useful
patterns in the data and generates a default probability to gauge the suitability of the applicant
for a loan. The whole underwriting process can be done using the mobile phone very quickly,
even within 10 minutes. The granted loan can be transferred to the applicant’s bank account
or e-wallets within 15 minutes which enhances customer experience.
Similarly, insurance underwriting can also leverage on an AI system to extract and analyze
data containing information on the applicant’s hazard likelihood, historical claims, life
attitude, and so on to generate an insurance quote quickly.
The AI underwriting system is superior as it can evaluate the borrowers without
loan/insurance histories that are essential in traditional underwriting. Compared to traditional
underwriting, banks and insurance companies will benefit from AI-based underwriting as it
reduces employee costs and may also lead to a decrease in fraud cases as the social media
data comes from the applicant’s real life activities.

You might also like