The safety of AI? The Bletchley Park statement is crucial

In detail the important initiative sanctioned by the main countries protagonists of global progress in the field of Artificial Intelligence

AI Security: World leaders gather at the 2023 Artificial Intelligence Security Summit at Bletchley Park in Britain
World leaders gathered at the 2023 Artificial Intelligence Security Summit at Bletchley Park in Britain (Photo: Leon Neal/Reuters)

Last November 1st and 2nd, during the AI Security Summit, held in the United Kingdom at Bletchley Park, in the County of Buckinghamshire, 28 countries have signed a declaration of intent regarding the development of Artificial Intelligence in safe conditions.

AI: the war that is about to break out will not be as we expect
SwissGPT: the Swiss AI that revolutionizes corporate security
Photogallery, the future of digital transformation in Liechtenstein

A location with a high symbolic value, for better or for worse…

The location where the summit was held, Bletchley Park, was the decryption center during the Second World War where Alan Turing and his team worked on the creation of “Christopher”, the computing machine that managed to undermine the code of Enigma, a German device used to encode secret military messages until then considered mathematically unassailable.

A place with a particularly high symbolic value, as it has been calculated that thanks to the English scholar the war lasted two years less, saving at least 14 million lives.

Unfortunately, the same place also symbolizes the fundamental indifference and ingratitude of many governments towards their saviors, given that Alan Turing committed suicide in 1954 after a year of "chemical castration" imposed on him for his homosexuality, with the grotesque ending of the royal "posthumous pardon". in 2013, as if he were the one who had to make amends for something.

All this also leaves us wondering about the current approach of the highest authorities towards the problems opened by the recent "explosion" of technologies linked to Artificial Intelligence.

Switzerland and the USA aligned on cybersecurity and digital technology
Water, grass and humanity: the cognitive limits of Artificial Intelligence
Artificial Intelligence also for the development of new drugs

Michelle Donelan visits Bletchley Park
Michelle Donelan, Secretary of State for Science, Innovation and Technology for the UK Government, visiting Bletchley Park during the AI ​​Safety Summit

But let's see what the declaration of intent contains

After a summary analysis of the opportunities for improvement offered by AI in the various sectors of human activities and the acknowledgment that the technologies linked to it are now widespread, the need for their safe development and to maintain always putting the needs of our species at the center.

A series of possible risks are then listed, some of which have already been verified in the field, linked to the use of AI systems, and in particular: risks inherent to human rights (to work, to care, to justice); risks inherent to the defense of privacy; risks linked to correct information (extremely credible misleading content); risks linked to technological "singularities" linked to AI (new disruptive technologies in sensitive contexts such as biotechnology and information technology).

In Switzerland the Federal Polytechnics for transparent and reliable AI
Artificial Intelligence is (also) a revolution in photo editing
Innovation and journalism: an often difficult coexistence

Tom Cruise deepfake example
An example of "deepfake", or the use of artificial intelligence techniques to make a person appear similar to the real actor Tom Cruise

Managing the problem beyond state borders

Since many of these risks are intrinsically international in nature, maximum cooperation between States is recommended to address them as they arise, as well as active participation of all actors starting from the largest and most important companies, therefore capable of " train” the most powerful AI, in a continuous, open and transparent comparison.

A network of collaborations will therefore be created at a global level with the aim of identifying the risks of artificial intelligence through a global and shared scientific approach to better understand its impact on society and consequently build common policies to address the related changes.

Michael von Liechtenstein on geopolitics and artificial intelligence
When old age is the dividing line between good and bad technology
The Big Data leitmotif of the “Swiss Statistics Days 2023”

The risks of the top-down approach in managing the topic

Several industry experts point out that, beyond the good intentions and declarations of wanting to continue open discussion with all the players, such a "top-down" approach risks "tying the hands" only of the large companies in the sector, instead leaving the smaller and therefore difficult to control entities free to all.

These are realities that will spread more and more precisely thanks to those technologies, currently developed by companies such as Google and OpenAI, which will allow for greater personalization of AI systems thanks to the optimization of the necessary resources, perhaps operating at the limits of legality.

In short, as the ancient Chinese curse says, interesting times are ahead.

That permanent leap of faith in Christopher Nolan's cinema
A roadmap to promote Liechtenstein as a digital hub
In Frauenfeld, AI is driving a new PET sorting plant

What really happened at the 2023 Artificial Intelligence Security Summit at Bletchley Park

The opening plenary of the 2023 Artificial Intelligence Security Summit at Bletchley Park

The closing plenary of the 2023 Artificial Intelligence Security Summit at Bletchley Park

The message of King Charles III of Great Britain to the public and the organizers of the AI ​​Safety Summit 2023

AI Safety : In a scene from the biopic
In a scene from the biopic "The Imitation Game," actor Benedict Cumberbatch plays Alan Turing alongside "Christopher," the computing machine he invented that helped defeat the Nazis