Why did Henry Kissinger co-author a book on artificial intelligence at the age of 98?

The former US secretary of state, national security advisor and historian has not yet excelled in the field of information technology - but artificial intelligence (AI) has evidently worried him for some time.

The occasion was a conference a few years ago, at which an employee of the AI ​​developer Deepmind, which belongs to Alphabet (Google), explained the now prominent program that outperformed the best human players in the traditional board game Go and other programs in chess that outperformed human players , quickly overshadowed.

Alexander Armbruster

Responsible editor for business online.

  • Follow I follow

Kissinger was stunned by the learning process behind it and soon thereafter published an article on AI in The Atlantic magazine under the dramatic headline "How the Enlightenment Ends", in which he raised some fears that were exaggerated from a professional point of view, but struck a nerve with the central question: What does it mean when knowledge and skills are created that sometimes come close to or surpass human abilities and skills, but cannot or cannot fully be understood by humans themselves?

From Socrates to BERT

This was followed by an exchange, in particular, with two technically and entrepreneurially experienced computer scientists, the former long-standing Google CEO Eric Schmidt and MIT professor Daniel Huttenlocher, who had previously researched at the Californian inventor laboratory Xerox Palo Alto Research Center (PARC).

For example, the three wondered how AI could affect health and medicine, space exploration, quantum physics, children's toys or warfare.

And they also dealt with possible more abstract consequences.

One conclusion reads: “To a certain extent, the three of us differ on how optimistic we should be about AI.

But we agree that this technology transforms human thought, knowledge,

They have now placed their discussions and assessments in an entertaining and stimulating book.

Its title, “The Age of AI”, is a judgment in itself: it elevates AI to the rank of other key technologies that have fundamentally changed human societies.

What the AI ​​can already do is demonstrated in programs such as AlphaZero and GPT-3, the historical review extends to Alan Turing's ideas.

The different approaches to AI are explained: rule-based approaches and the learning algorithms based on enormous amounts of data and computing capacities, which have made many AI successes possible in the recent past;

and not only in a playful way, but also with commercially relevant developments.

The balance of power between states

In addition, the contributors are designing a philosophical course that ranges from Socrates to Descartes and Kant to BERT, an artificial neural network that is intended to further develop the Google search engine.

Last but not least, it is about the fact that the cognitive capacities of humans are a decisive reference, but their rank is not fixed.

This may sound banal and, when it comes to AI, is not necessarily the most urgent task for many AI experts.

However, it confronts the reader with the original claim of the AI ​​masterminds after the Second World War, who did not just have in mind a chess program unbeatable by humans or a search engine, but stated the goal of functionally replicating the human brain.

In a second section of the book, Kissinger, Schmidt and Huttenlocher span the gap beyond the technical, philosophical and economic aspects to the hard-hitting power politics.

The AI ​​has actually arrived there a long time ago, foreign and security politicians and the military are dealing with it, the question of who leads in the AI ​​is relevant for the balance of power between states and the United States, China, Russia and the European Union are bringing their own Initiatives are underway to ensure that we do not fall behind.

One of the good things about this book is that the AI ​​is actually looked at from different perspectives without - as sometimes happens - straying into Hollywood-like.

There should be more such representations.

Henry A. Kissinger, Eric Schmidt and Daniel Huttenlocher: "The Age of AI".

And Our Human Future.

Little, Brown and Company, New York 2021. 272 ​​pp., hardcover, €27.99.