Artificial intelligence (AI) has become a ubiquitous part of our lives. From personalized recommendations to medical diagnoses, AI has been shaping the world around us for years. Despite its omnipresence, AI continues to remain, if not entirely in a black box, at least in a shadowy one. ISACA’s Artificial Intelligence: A Primer on Machine Learning, Deep Learning, and Neural Networks seeks to illuminate that box with a clear and accessible guide, comprehensively covering the field with a generous offering of practical use cases, examples and illustrations.
“At its core, AI is an umbrella term that is not uniformly understood. This publication arose from ISACA’s desire to provide essential information to digital trust practitioners, regardless of work function, and ultimately became a body of knowledge,” said Jon Brandt, ISACA’s Director, Professional Practices and Innovation. “The primer is purposely comprehensive to increase the readers’ knowledge of AI and its inner workings, regardless of their level of familiarity with AI.”
With a goal of appealing to those just beginning on their journeys in AI and those with some knowledge in the field, the publication offers a structured approach for readers to advance their understanding of AI by, first, introducing fundamental AI concepts and the types of AI before progressing further into machine learning (ML), large language models (LLMs), neural networks (NNs), deep learning (DL), generative artificial intelligence (Generative AI) and AI’s future.
The immense promise of AI and the pace at which it has been incorporated into popular applications indicates a crucial need to address its potential risks thoughtfully and comprehensively, ensuring responsible development and deployment of applications driven by machine intelligence. Artificial Intelligence: A Primer on Machine Learning, Deep Learning, and Neural Networks addresses this concern with an entire chapter, arguably one of the most important in the book, examining “the urgent need for responsible AI governance and a new regulatory authority dedicated to AI.”1
As the use of AI continues, so, too, grows the risk of cyberattacks and unintended use, requiring a concerted effort to effectively mitigate the multiple risks associated with it. The timely coverage of organizational governance, legislative actions and the development of frameworks for responsible AI equips readers with knowledge needed to navigate the complexities of the ethical and societal landscape of AI that has been evolving since the 1940s.
Tracing AI’s evolution from early theoretical concepts to the powerful applications we use today, the primer covers groundbreaking events in the advancement of machine intelligence, ranging from the development of ELIZA, “an attempt to simulate human-like conversation,”2 to the mainstream media attention given the historic chess match between IBM’s supercomputer Deep Blue and the reigning human chess champion, Garry Kasparov, to recent developments in the subfield of Generative AI, which use generative adversarial networks (GANs) to output new content across media formats.
As for that “black box” referred to earlier, Artificial Intelligence: A Primer on Machine Learning, Deep Learning, and Neural Networks examines the different types of AI that one might encounter categorized by functionality and capabilities. Whether reading the book from cover to cover or using it as a reference resource, readers will take from it knowledge covering traditional analytics and predictive AI, to emerging AI capabilities such as explainable AI (XAI) that make AI processes and outputs increasingly transparent and trustworthy.
A good portion of the book is reserved, and rightfully so, for machine learning (ML), a subfield of modern AI: “ML is an innovative subfield of AI that employs models and algorithms to enable machines to learn from data without explicit programming. These models and algorithms are based on various mathematical concepts such as statistics, probability, linear algebra, calculus, and optimization. As an essential component of data science, ML leverages advanced statistical methods to develop robust algorithms capable of solving complex problems.”3
Emphasizing the crucial role of data preparation in machine learning, the book covers various types of datasets and data processes, from data collection to data splitting for training, validation, and testing, and extensively explores supervised, unsupervised, and reinforcement learning, along with the various techniques each uses to provide their outputs across various domains. It also examines neural networks, the foundation of DL – a powerful subset of machine learning that utilizes complex architectures to tackle increasingly intricate tasks. Various DL architectures, such as autoencoders, convolutional neural networks (CNNs), recurrent neural networks (RNNs), and transformers, are described, with particular attention given to their specific strengths and weaknesses.
Recognizing that AI is not a singular field of study but rather a multi-disciplinary one, addressing its technical and ethical dimensions, this publication is designed for anyone engaged in using AI and monitoring and challenging its outputs; anyone exploring how machines learn and make decisions as well as generate human-like content; and anyone looking forward to what the future holds as AI begins to understand and respond to human emotions and social cues. Offering clear and accessible explanations, Artificial Intelligence: A Primer on Machine Learning, Deep Learning, and Neural Networks empowers its readers to engage in the current and future discourse on AI.
1 ISACA, Artificial Intelligence: A Primer on Machine Learning, Deep Learning, and Neural Networks, p. 137, 2024.
2 IBID., p. 15
3 IBID., p. 29