Machina sapiens vs Homo sapiens?
A book to make us consider the great issues raised by artificial intelligence and our future has just been published.
Some are convinced we are truly on the cusp of machines that genuinely think (while others continue to nurture considerable doubts). These machines, in any case, though created by humans could soon not only replace humanity in performing many tasks but represent alternative creatures. Never before has the world – economic and otherwise – been so divided on the topic of artificial intelligence, raising many doubts flanked by just as many dreams. Now as never before, it’s necessary to understand. Reading Machina sapiens. L’algoritmo che ci ha rubato il segreto della conoscenza (Machina sapiens: the algorithm that stole our secret of knowledge) by Nello Cristianini, who lectures in Artificial Intelligence at the University of Bath, certainly aids understanding, getting more of a (precise) idea and above all reasoning with a firm foundation of knowledge.
In just a few pages, Cristianini manages to provide the entire basic toolbox required to deal shrewdly with a topic that is complex and constantly evolving, starting from a consideration: machine intelligence arises from interaction between a mathematical mechanism and an extraordinary quantity of texts “that no one has ever tried to connect and distil before”. The consequence of all this is not just a language but a “model of the world with capacities that are still unexplored and unexplained”. Every unknown arises from this situation. This is where the great fascination of intelligent machines lies, but also the whole complex of doubts that they generate and which the book tries at least to put in order, if not to resolve.
To address the topic, Cristianini (rightly) starts from Alan Turing, effectively the father of computer science, who in the 1950s asked himself whether machines could think. From there the author comes to consider what is happening today, that new intelligent agents such as ChatGPT have proven capable of carrying out tasks that go far beyond the initial intentions of their creators. As Cristianini emphasises, this is because while these machines were trained for certain abilities, others emerged spontaneously as they read thousand of books and millions of web pages.
The narrative is based around three groups: scientists (who designed and created the machines), people (who are starting to deal with these machines in everyday life), and the machines themselves (increasingly revealing themselves as intelligent beings). In the epilogue, Cristianini doesn’t present conclusions but questions and suggestions. One relates to reaching a so-called “critical mass”, “the possibility that there is a size threshold beyond which the performance of the intelligent machine begins to accelerate”. Another relates to the fact that, once begun, the path that intelligent machines are capable of taking would quickly lead them to “outperform the weakness of our capabilities”. Cristianini then speaks of “emergent abilities” which confront humanity with the question of what could happen to it tomorrow (in terms of economics and work as well as socially). Again in the epilogue, the author calls the decisive role of rules to mind, and thus of wise policy also when faced with machines capable of thinking.
Nello Cristianini’s book doesn’t offer certainties or even definitive answers, and this is precisely where its great value lies: it helps the reader to think (intelligent machines notwithstanding).
Machina sapiens. L’algoritmo che ci ha rubato il segreto della conoscenza
Nello Cristianini
il Mulino, 2024
A book to make us consider the great issues raised by artificial intelligence and our future has just been published.
Some are convinced we are truly on the cusp of machines that genuinely think (while others continue to nurture considerable doubts). These machines, in any case, though created by humans could soon not only replace humanity in performing many tasks but represent alternative creatures. Never before has the world – economic and otherwise – been so divided on the topic of artificial intelligence, raising many doubts flanked by just as many dreams. Now as never before, it’s necessary to understand. Reading Machina sapiens. L’algoritmo che ci ha rubato il segreto della conoscenza (Machina sapiens: the algorithm that stole our secret of knowledge) by Nello Cristianini, who lectures in Artificial Intelligence at the University of Bath, certainly aids understanding, getting more of a (precise) idea and above all reasoning with a firm foundation of knowledge.
In just a few pages, Cristianini manages to provide the entire basic toolbox required to deal shrewdly with a topic that is complex and constantly evolving, starting from a consideration: machine intelligence arises from interaction between a mathematical mechanism and an extraordinary quantity of texts “that no one has ever tried to connect and distil before”. The consequence of all this is not just a language but a “model of the world with capacities that are still unexplored and unexplained”. Every unknown arises from this situation. This is where the great fascination of intelligent machines lies, but also the whole complex of doubts that they generate and which the book tries at least to put in order, if not to resolve.
To address the topic, Cristianini (rightly) starts from Alan Turing, effectively the father of computer science, who in the 1950s asked himself whether machines could think. From there the author comes to consider what is happening today, that new intelligent agents such as ChatGPT have proven capable of carrying out tasks that go far beyond the initial intentions of their creators. As Cristianini emphasises, this is because while these machines were trained for certain abilities, others emerged spontaneously as they read thousand of books and millions of web pages.
The narrative is based around three groups: scientists (who designed and created the machines), people (who are starting to deal with these machines in everyday life), and the machines themselves (increasingly revealing themselves as intelligent beings). In the epilogue, Cristianini doesn’t present conclusions but questions and suggestions. One relates to reaching a so-called “critical mass”, “the possibility that there is a size threshold beyond which the performance of the intelligent machine begins to accelerate”. Another relates to the fact that, once begun, the path that intelligent machines are capable of taking would quickly lead them to “outperform the weakness of our capabilities”. Cristianini then speaks of “emergent abilities” which confront humanity with the question of what could happen to it tomorrow (in terms of economics and work as well as socially). Again in the epilogue, the author calls the decisive role of rules to mind, and thus of wise policy also when faced with machines capable of thinking.
Nello Cristianini’s book doesn’t offer certainties or even definitive answers, and this is precisely where its great value lies: it helps the reader to think (intelligent machines notwithstanding).
Machina sapiens. L’algoritmo che ci ha rubato il segreto della conoscenza
Nello Cristianini
il Mulino, 2024