By Cezzy Kanionga, Senior Consultant
Artificial intelligence, or AI, is a program that seeks to mimic human intelligence through computational algorithms. Its creation allows computers to perform operations and think like a human being. Since 2010, the development of artificial intelligence has been accelerated by big data. In this article, largely drawn from the Datascientest publication (https://datascientest.com/intelligence-artificielle-definition), you will not only discover what you need to know about AI, but you will also learn about the international standard for AI management systems recently published by the International Organization for Standardization (ISO).
Table of Contents
- Definition of Artificial Intelligence
- The History of Artificial Intelligence
- What are the Types of Artificial Intelligence?
- What is Artificial Intelligence Used For?
- Risks and Challenges Associated with AI
- What Types of Standards Does ISO Offer for Artificial Intelligence?
- What is an Artificial Intelligence Management System?
- What are the Objectives of ISO/IEC 42001?
- Why is ISO/IEC 42001 Essential?
- What are the Main Benefits of Implementing ISO/IEC 42001?
- Who is ISO/IEC 42001 For?
- Zylloo Consult and PECB Certifications in AI
- Definition of Artificial Intelligence (AI)
According to John McCarthy, one of the pioneers in the field, AI is “the science and engineering of making intelligent machines.” Artificial intelligence (AI) is a branch of computer science that seeks to create systems capable of performing tasks that would normally require human intelligence.
However, AI is often considered a broad and multidimensional concept, difficult to define precisely due to its expansive and constantly evolving nature. For example, technologies ranging from simple recommendation algorithms used by Netflix, according to Forbes, to complex autonomous driving systems developed by companies like Tesla, are all considered AI. This diversity makes the field both fascinating and enigmatic, with definitions evolving as the technology advances.
The History of Artificial Intelligence
The history of artificial intelligence begins in 1943 with the publication of the article “A Logical Calculus of Ideas Immanent in Nervous Activity” by Warren McCullough and Walter Pitts. In this paper, the scientists presented the first mathematical model for creating a neural network.
In 1950, Snarc, the first neural network computer, was created by two Harvard students: Marvin Minsky and Dean Edmonds. That same year, Alan Turing published the Turing Test, which is still used to evaluate AI. From this test stem the foundations of artificial intelligence, its vision, and its objectives: to replicate or simulate human intelligence in machines.
But it wasn’t until 1956 that the term “artificial intelligence” was first uttered, during John McCarthy’s “Dartmouth Summer Research Project on Artificial Intelligence” conference. At this event, researchers presented the goals and vision of AI. Many consider this conference to be the true birth of artificial intelligence as we know it.
Several years passed, and work on artificial intelligence continued. In 1959, Arthur Samuel coined the term “Machine Learning” (ML) while working at IBM. In 1989, the Frenchman Yann LeCun developed the first neural network capable of recognizing handwritten numbers; this invention would lead to the development of “deep learning.”
And ten years later, in 1997, a major event marked the history of AI. IBM’s Deep Blue system defeated world chess champion Garry Kasparov. For the first time, the machine has defeated Man.
What are the types of artificial intelligence?
Thanks to the development of artificial intelligence and technologies like deep learning and machine learning, researchers generally agree on three types of artificial intelligence:
General Artificial Intelligence
General or deep AI is an artificial intelligence capable of performing any cognitive task, much like a human or animal. While still considered hypothetical, some scientists are questioning GPT-4 and the likelihood that it represents an early form of general AI. In this regard, a large majority of AI researchers believe that humanity possesses the necessary technology to create general AI, particularly through neural networks.
Strong Artificial Intelligence
Strong AI, or superintelligence, refers to a model that references philosophical knowledge and shows signs of independent consciousness. While it may seem like a science fiction scenario, AI researchers believe that strong AI is currently impossible to create. They argue that the concept of consciousness and feelings cannot exist in mathematical systems that manipulate and respond using symbols and calculations.
Weak Artificial Intelligence
The final distinction within artificial intelligence is weak or narrow AI. This type of AI is a system capable of performing a single task almost perfectly, without the need for human supervision. It is the most widely used model and is designed to accelerate various processes across different industries.
What is the purpose of Artificial Intelligence?
Artificial intelligence has transformed numerous sectors. These include medicine, science, finance, the automotive industry, and many others.
In medicine, AI is used to diagnose and predict diseases, enabling early detection and rapid intervention. It is also applied in pharmaceutical research to accelerate drug discovery and improve treatments.
In scientific research, AI analyzes vast amounts of data and makes discoveries in fields such as astrophysics, genomics, biology, and chemistry. It accelerates scientific progress and opens up new research avenues.
With the arrival of GPT-3.5 in November 2022, a powerful LLM (large language model), the potential of artificial intelligence has increased tenfold. AI is now also used in creative fields such as text and image generation, and even in audiovisual production, thanks to applications like VALL-E, Midjouney, and GEN-2.
Risks and challenges related to AI
Faced with the massive changes brought about by artificial intelligence, many people are also beginning to question the dangers it could pose. Considered as dangerous as the nuclear bomb by business leaders like Sam Altman, CEO of OpenAI, and by researchers, AI could become a threat to humanity.
Thanks to its capacity for learning and evolving, AI could one day surpass humans in certain fields, and thus replace them. This is certainly the conclusion of a Goldman Sachs study, which claims that, within a few years, more than 300 million jobs will disappear. At the same time, people are questioning their privacy and personal data. Needing data to improve itself, will AI gain access to everyone’s personal data under the guise of major technological and economic development? Concerned about this catastrophic scenario, Europe and other companies, such as Google, are implementing regulations within their systems or through legislation like the AI Act.
But AI isn’t the only potential danger; everything also depends on how it’s used. Even seemingly harmless AI could be misused maliciously. We can already see this with the rise of “DeepFakes”: fabricated videos created using deep learning to portray a person in a compromising situation.
Artificial intelligence will continue to develop rapidly in the coming years. It is now up to humanity to regulate it in order to develop healthy and ethical AI. This is where the implementation of management system standards that allow for agreed-upon management of AI development and use becomes essential.
What types of standards does ISO propose for artificial intelligence?
The International Organization for Standardization (ISO) has published several standards that help mitigate risks and maximize the benefits of AI, including ISO/IEC 22989, which establishes AI terminology and describes AI concepts; ISO/IEC 23053, which provides a framework for AI and machine learning (ML) to describe a generic AI system using ML technology; and ISO/IEC 23894, which offers recommendations for managing AI risk within an organization.
ISO/IEC 42001, on the other hand, is a management system standard (MSS). Implementing this standard involves establishing policies and procedures for the proper governance of AI within an organization, based on the PDCA (Plan-Do-Check-Act) cycle. Rather than focusing in detail on specific AI applications, this standard offers a practical approach to managing the risks and opportunities associated with AI within an organization. It proves to be an asset for any company or entity.
What is an artificial intelligence management system and what are the objectives of ISO/IEC 42001?
An AI management system, as defined in ISO/IEC 42001, is based on a set of interrelated or interacting elements within an organization, used to establish policies and objectives, as well as processes to support the achievement of those objectives, in order to develop, deliver, or use AI systems responsibly.
ISO/IEC 42001 specifies requirements and recommendations for establishing, implementing, maintaining, and continually improving an AI management system (AIMS) tailored to an organization’s context. It provides organizations with detailed guidelines for using AI effectively and responsibly, while taking into account the rapid evolution of this technology. Designed to cover the various aspects of artificial intelligence and the diverse applications an organization may use, this standard offers an integrated approach to managing AI projects, from risk assessment to effective risk mitigation. It is therefore designed to enable organizations offering or using products or services that require AI to ensure the development and use of AI systems in a responsible manner.
Why is ISO/IEC 42001 essential?
ISO/IEC 42001, the world’s first AI management system standard, offers particularly useful guidelines in a rapidly evolving technological field. It addresses the unique challenges posed by AI, including ethical considerations, transparency, and continuous learning. It provides organizations with a structured methodology for managing the risks and opportunities associated with AI, while balancing innovation and governance.
What are the main advantages of implementing ISO/IEC 42001?
Responsible AI: Ensuring the ethical and responsible use of artificial intelligence.
Reputation management: Building trust in AI applications.
AI governance: Supporting compliance with legal and regulatory requirements.
Practical advice: Effectively managing AI-specific risks.
Identifying opportunities: Fostering innovation within a structured framework.
Who is ISO/IEC 42001 intended for?
This standard is designed for organizations of all sizes that develop, provide, or use products or services that utilize AI. Applicable to all sectors, it is equally relevant for public sector organizations, businesses, and non-profit organizations.
Zylloo Consult and PECB certifications in AI
Zylloo Consult is one of the distributors in Africa of training and certification services for individual skills and organizational management systems from the Canadian firm PECB. Zylloo Consult organizes PECB training events in self-study, online, and in-person formats. Customized training modules tailored to partners’ needs are also available upon request.
The ISO/IEC 4001:2013 standard had just been published in December 2023, and only in English. Training and certification courses in this area are available in English for the PECB LEAD IMPLEMENTER and PECB LEAD AUDITOR certifications. While awaiting the French version of the standard and the development of training materials in French, we invite our partners interested in mastering AI and its effective management to contact us for guidance on its use.