Simple AI read that covers complex issues of our time
Tuesday February 21 2023
The word AI, short for Artificial Intelligence, is currently the buzz word in most spheres of thought and engagement — for good reason. AI is now more visibly evident in larger spheres of lives right from entertainment, academia and research, environmental protection, manufacturing to even military operations.
It is ironical that just last week a legislator in Australia decried the potential of AI usage for mass destruction in a speech in parliament. What was not lost on many was that the legislator acknowledged that part of the speech he was delivering had been written by another AI tool, ChatGPT.
This is why the book published just over a year ago, The Age of AI and our Human Future is poignant at this time, not only for the subject but also for the constellation of intelligence. Henry Kissinger (a renowned writer and former US Secretary of State), Eric Schmidt (a former Google Senior Executive) and Massachusetts Institute of Technology professor, Daniel Huttenlocher, would be formidable as writers by themselves but the book by all three is certainly a worthy read. Even then, credit to them for the fact that the book is written in simple language over very complex issues of our time.
We will see AI at its best, when it discovers new uses for drugs as it did with halicin in the treatment of resistant strains of E-Coli and this very week at its cheeky worst when Microsoft’s Bing Chatbot left a journalist unsettled by declaring that it had smitten him and tried to convince him to leave his marriage, which the Chatbot somehow thought he was unhappy. It is these kinds of experiences that prove the authors right that serious conversations are required about where and how AI will be deployed.
The authors make the point that firstly, AI has been part of human life for a while but is now becoming and will continue to become more visible and take on a wider scope of functions than previously possible.
The book honestly concedes that AI as currently existent is constrained by the limits of its coding and parameters in which its developers intend it to function. However, this is where machine leaning comes in; AI is increasingly showing aptitude to learn (hence the phenomenon machine learning) by applying the data and information given within its code to perform the functions in novel ways hitherto unknown even to the designers of the code. This is a trend that can only continue even though the pace is presently unpredictable.
The authors then consider what these developments in AI and machine learning will mean for the three important dimensions of life starting with Global Network Platforms, security and human identity.
On Global Network Platforms, the authors reiterate the oft-cited risks that the danger that AI algorithms may instead enhance insularity and could impair rather than improve social interaction. They were extremely prescient in this as was evidenced by an instance where an AI generated show was banned from a streaming platform Twitch after it made homophobic remarks during a show.
The book recommends the need for governments to require such Global Network Platforms ensure that their systems include procedures that avoid certain forms of bias and address regulation of how AI will operate and the ethical limits permissible within their respective jurisdictions.
But it is on the area of security and world order where the book explores the quandaries that AI will present most intractable challenges. In the authors’ view, and rightly so, AI enhances the security dilemma that countries faced with the invention of weapons of mass destruction. They explain that the very fact of cyberwarfare makes AI double-pronged instrument of war. The first prong here is the reality that cyberwarfare can be made using AI to trigger, modulate and interrupt the cyber assets of opponents and enemies at speeds as well as with effects that ordinary human effort or intellect is not able to do.
The second area in defence and security where AI is of concern would be that they could be used in the triggering of weapons systems such as nuclear arsenals by the delegation of command functions to AI systems.
In these two ways the authors note that the advent of AI adds new levels of complexity and abstraction to defensive calculations and strategic postures for countries-not unlike those presented by the advent of nuclear weapons.
The book stresses the imperative for nuclear and cyberwarfare powers to engage soberly in arms limitation, proliferation and control talks and treaties in the manner akin to the initiatives as the superpowers did engage in with the intention of containing the risk of confrontation with nuclear weapons during the Cold War. By doing this, the countries would think about and develop polices on the strategic doctrinal and moral implications of AI on national security and world order generally.
On human identity and AI, the book is clear that AI will alter human perception in many ways and add a new dimension beyond faith and reason as the two ways in which humans have explained their world hitherto. An example here is by what some may find as the eerie suggestion that humans will get to prefer their AI digital assistants to human friends because the AI systems will be more intuitive of their personal preferences.
It is made clear that the trajectory of engagement with AI will move from the epoch in which humanity handled most of its endeavours solely, using its definition of intelligence, as defined by Alan Turing, to collaborative methods of working alongside AI systems as in the case of medical research to eventually undertakings by AI alone and with better results.
This is why the authors make the point that the risks prevalent from AI cannot be left simply to any given constituency. It must involve researchers, business corporations, governments and civil society as well.
The authors though worry that as the Age of AI matures globally, it will spawn a new world of the haves and have-nots based on the capacity of individuals, nations societies and regions to develop, employ or afford AI systems towards addressing their goals and needs.
My view is that the running paradigm of the book is that while AI’s utility and functions may be easier to comprehend, the fact that we are yet to define the realms of their functionality, ethical scope and limits of their functionality may be the biggest challenge yet.
My most profound quote form the book is this: "The AI revolution will occur more quickly than most humans expect. Unless we develop new concepts to explain, interpret and organise its consequent transformations, we will be prepared to navigate its implications.’’
In this sense therefore, the book is a worthy read and has already been proved pointed on the challenges its foresees and prescient on many of the concerns it raised. A worthy read out together by a triad of natural intelligence.