Let Us Agree on One Definition for AI

17 Jul 2023
Dr. Talal Abu-Ghazaleh share Jumana Abu-Ghazaleh’s views on Artificial Intelligence

In 2002, I asked Bill Gates at the Microsoft Executive Summit Meeting in Paris, what the next era would be, to which he replied the era of Artificial Intelligence. With the AI era well and truly upon us, I have been avidly following this technology with my daughter Jumana Abu-Ghazaleh, and sharing her knowledge as a keen observer of this space.
Today, everyone is talking about AI and its potential to change and revolutionize our world in a myriad of different ways. It has become a vibrant topic that sparks diverse and intense discussions about its pros and cons, its reality and fantasy, its biases and ethics, as well as its threats and promises. With ChatGPT and AI going mainstream,the general public the general public can now also see its potential, dangers and limits.
Having assessed what is going on in the world of AI and having consulted with Jumana and      others, what seems to be missing is a clear, universally accepted definition of AI. It seems to mean different things to different people: The US Department of State and the OECD refer to it as a kind of system, the US Congress has called it a computerized system and OpenAI has labelled it a highly autonomous system.
Meanwhile, companies like Microsoft say that it’s a capability and Google claims it’s a set of technologies. The EU says it’s a family of technologies, while China considers it a strategic technology. Meta has used various definitions over the years, enough to confuse ChatGPT into providing a definition and then retracting it.
With so many different takes on AI, there clearly seems to be a problem in defining what it is that we are talking about here. What exactly are we talking about when we refer to AI? If there is ambiguity about the thing we are referring to, how can we have meaningful discussions about how it is developed, governed, used and applied, or about its merits, dangers and its future?
As an accountant by trade and as the founder and chairman of one of the largest audit firms in the world, TAG.Global, I think precision and clarity about what things mean is something absolutely essential. In 2019, Jumana wrote a piece for ‘The Startup’ about the state of data-intensive and data-extractive technology that started with this line: “Everyone knows something’s wrong - they just can’t agree on what it is.” I could have started this article the same way.
Before we take AI discussion any further, we must settle on a definition of what it is that is both accurate and meaningful. As has been said in management circles “You cannot manage what you cannot measure”. I would extend this further and say, “You cannot measure what you have not defined”. If we are going to build a healthy, diverse and vibrant AI sector, we need to be clear from the get-go about our terminologies, definitions and hierarchies as these are the basis from which a holistic AI ecosystem can sprout.
We need a well-defined AI taxonomy with precise nomenclature before we proceed. We must have conversations about what AI is and what makes a system artificially intelligent.
To illustrate the quagmire we are in, we are already in a world that defines AI as multitude of things including a self-driving car, a facial recognition tool, a ChatBot, a business model, a robot, a language interpreter, a tool to help reduce climate change, a chess expert, an early disease detection assistant, and more.
I believe that it’s never too late to make things right and that its time to take meaningful action. I suggest we form a commission -call it the Commission for AI in the Public Interest (CAIPI)- comprised of experts in various related fields, including data science, law, ethics, neuroscience, philosophy, and others. They should be tasked with stress-testing and standardizing the definition so that those working in the field, as well as those using and being used by the field, are clear on the rules of engagement and can expect a modicum of consistency and accountability. It’s called governance and it’s part of literally every other aspect of our life.
Being the previous Chair of the Advisory Committee on Internet Governance for the United Nations Information and Communication Technologies Task Force and having hosted numerous international forums on governance -as well as having published numerous works on governance- I invite Bill Gates and others to join me in taking action to clearly define AI. I will happily host such a gathering at Talal Abu-Ghazaleh Academy and at Talal Abu-Ghazaleh Digital University, where we are actively investigating, exploring, and teaching various aspects of AI.
Talal Abu-Ghazaleh