Page 12 - AI for Good - Impact Report

AI for Good

predefined rules. Deep learning (DL) is a subset of ML that leverages artificial neural networks
(ANNs) with multiple hidden layers – hence “deep” -, enabling the processing of complex data
inputs such as images or speech. The newest and most widely recognized area of AI is called
Generative AI (GenAI). It represents a specialized application of deep learning where ANNs can
also generate realistic content. GenAI applications of today are based on foundation models.
Foundation models are a class of AI models trained on broad datasets and are designed to be
adaptable to a wide range of downstream tasks, such as image and video generation, image
description, or translation. Most of these models are built on the transformer architecture,
which has proven highly effective in natural language processing (NLP) tasks. The versatility of
foundation models extends to various other domains or modalities, including computer vision,
speech recognition, and even generative tasks, where they can produce new content.
9
Since the public GenAI race began in November 2022 with the release of OpenAI’s “ChatGPT”,
organizations have increasingly focused on using its potential. The ability to create new
10
content, such as text, images, and videos, in response to user queries has made it popular
in sectors like healthcare, finance, and entertainment. Numerous tools have demonstrated
practical applications of GenAI, from simulating human conversations to aiding in software
development and generating digital images from textual descriptions. While GenAI currently
11
receives the most attention, other areas of AI are also being developed and finding their way
into different applications.

Various industries are adopting AI technologies

For companies, AI adoption is increasingly vital, with 94% of global business leaders viewing AI
as critical for their organization’s success in the next five years. Notably, 67% of organizations
12
are ramping up investments in GenAI due to its demonstrated value. Global AI market revenue
13
is projected to grow by a 19% Compound Annual Growth Rate (CAGR) over the next decade,
surpassing US$2 trillion by 2031. Customized AI technologies are already catering to the
14
specific needs of various industries, with significant potential to drive positive impact in sectors
such as healthcare, education, energy, agriculture, and finance. The following section provides
a closer look at these industries, featuring typical use cases for illustrative purposes.

In healthcare, developments in AI and machine learning are driving innovation and transforming
operating models by enabling personalized patient care, predictive analytics for disease
prevention, and efficient management of healthcare resources.
In the long run, AI in healthcare is shifting the focus from treating diseases to early diagnosis
and prevention. This shift is driven by advanced computing power and smart algorithms
15
that can identify patterns in digital data and images, making diagnosis and treatment more
data-driven. For example, a US-based medical imaging startup uses large learning models
for early disease detection in stroke care, cardiology, and oncology. Similarly, another startup
founded in Belgium, has developed an AI solution, using deep learning algorithms for early
detection of traumatic brain injuries. In Singapore, a cloud-based platform for diagnosing retinal
conditions such as diabetic retinopathy, macular degeneration, and glaucoma was deployed,
which can help address the scarcity of eye specialists in rural areas. These advancements can
16
help enable healthcare providers to identify diseases early, improving patient outcomes and
reducing healthcare costs.

AI is significantly accelerating the drug discovery process by conducting numerous experiments
and analyzing vast datasets to identify new medications. Pharmaceutical companies use AI to
17

2