Page 721 - AI for Good Innovate for Impact

AI for Good Innovate for Impact

(continued)

Item Details
Code reposi- N/A
tories cities 4.8: Smart home/

2 Use Case Description

2�1 Description

Stable internet connections are necessary in order to partake in online meetings, work
interviews, internet courses, and online conferences. However, in less developed regions like
Africa, the constant and predominantly unforeseeable network crashes mean opportunities
go by for pupils, job applicants, and working professionals. Job interviewees miss valuable
interviews due to abrupt shutdowns, and pupils are interrupted midway through imperative
online exams. This disruption influences companies and virtual gatherings as well, leading to
missed productivity and revenues.

While there are some present mechanisms that might detect or even anticipate network failures,
their prime objective is to inform the service providers or carry out retrospective analysis. This
application, however, proposes an AI-based network quality prediction system which not only
predicts failures but also assists users in planning ahead for their online behaviour.

The system leverages a combination of past network history, real-time user feedback, weather
conditions, and telecom operator feeds to provide accurate, location-specific predictions
of network quality. Users can input a time and place to receive AI-based predictions of
network stability. Crucially, the platform goes beyond basic passive notifications by offering
recommendations on the optimal available time slots for online events, enabling users to
schedule with confidence.

It seamlessly integrates with popular applications such as Google Calendar, Zoom, and
Microsoft Teams, giving reminder notifications to reschedule or organize contingency options
in advance. This anticipation foresees and mitigates risks of network failures and provides more
efficient online experiences.
Unlike typical telecom coverage maps providing a general availability or speed test applications
that provide one-off measurements, this system provides dynamic, forward-looking information.
It shifts the paradigm from reactive to proactive management of connectivity. This kind of
proactive functionality is especially helpful in low-connectivity environments, where digital
inclusion is riding on the power to plan effectively beyond infrastructure limitations.
By maximizing the utilization of network resources in that and at which times they are accessed,
the system maximizes overall digital efficiency and allows users to avoid disruptions before
they occur. One discussed problem is that some data sets—particularly those that come from
telecom firms—may not be readily accessible for training, but federated data collection and
crowdsourcing methods eliminate this issue.

With real-time and forecast network quality data, the solution not only aims to detect failures but
to enable better digital planning, reduce lost opportunities, and enhance online experiences,
especially in areas prone to network instability.

685