La guerre des intelligences
This book is well worth a read for some insight into how artificial intelligence is going to affect our lives and the history of humanity. It poses all the most fundamental questions concerning the future of our societies. It supplies well-argued answers, not entirely lacking in objectivity, for some of these questions; it also supplies scenarios which, despite the author’s optimism, highlight the dangers stemming from the inevitable introduction of this technology.
It is also a book that no reader can be entirely indifferent to, so heavily impregnated is it with the author’s personal beliefs. Some of these beliefs border on obsession and need to many redundancies. The author, Laurent Alexandre, is a surgeon, neurobiologist and product of France’s most prestigious graduate college of administration, the ENA, the founder of the French digital platform Doctissimo and of several other high-tech companies, in Belgium and elsewhere. Some of his ethical choices, such as his generally favourable view of transhumanism, notwithstanding his criticism of some of its more extreme manifestations, or his rejection of punitive ecology, are debatable. In many cases, his rationale comes from a place of bias, despite being presented as logical and neutral. The process, presented as scientific, is the product of a line of argument that is frequently based on a compilation of insufficient or inadequate data. This is the case, for instance, when the author sets out to challenge the effects of climate change and again when his argument relies on estimating the number of victims of witch hunts, which were carried out for a period of more than two centuries, Aztec sacrifices, the estimated number of victims of which is well below that of the victims of the genocides of the 20th century alone, and an empirical evaluation of crime since the Middle Ages to assert that “the world has never been such a safe place” (our translation throughout). However, Alexandre also quite right points out that Europe is way behind the curve in new technologies and their geopolitical consequences.
“ChatGPT marks a new stage in the production of networks of artificial neurons. Its ability to stimulate the human spirit is astonishing, even though it still suffers from digital hallucinations”, Alexandre states, clarifying the nature of these “digital hallucinations” by explaining that “when it is unfamiliar with a given subject, ChatGPT makes it up with disconcerting aplomb”. This is highly reminiscent of much of human behaviour! The references made to “the human spirit”, which the Oxford English Dictionary defines as “the immaterial aspect of a person”, is quite strange, as it relates to a concept that is religious in its nature (in Judaeo-Christian theology, it is the breath of God) and cannot be reproduced artificially. What ChatGPT may manage to simulate is, logically, a form of human intelligence.
“We are going into a world in which the production of artificial intelligence will be infinite. The head of NVIDIA, the biggest producer of microprocessors for use in AI, predicts that in 10 years’ time, AI will be a million times more powerful than ChatGPT. Sam Altman predicts that the amount of intelligence on earth will double every eighteen months. Even worse, AI is becoming virtually free of charge while biological intelligence is rare, limited, takes a long time to produce, is extremely expensive, those on strike and is at odds with labour as a value. This difference automatically leads to a cognitive crisis”, the author continues, adding that “generative AI forms such as ChatGPT will do far more to structure human thought than search engines like Bing and Google. Controlling ChatGPT gives its owners, OpenAI and Microsoft, a huge amount of power”.
Alexandre goes on to point out that even the capricious billionaire Elon Musk “believes that one of the greatest threats posed by AI is its capacity to be manipulated by businesses” and is calling for it to be regulated, although this did not prevent him from launching X.AI on 14 April 2023 as a rival to ChatGPT. “An enormous planetary race is being run. Its aim is simple: to dominate the world via AI”, the author writes.
“Consumers – ourselves – are AI’s useful idiots. We are feeding the digital machine of tomorrow without realising it. We believe that our smartphones are the pinnacle of man’s technological superiority, not understanding that in reality, it is the tool of our enslavement. AI’s raw material is information. Where does this come from? From us, with our billions of Google searches and the billions of images we upload to Facebook every day. For ‘deep learning’, the avalanche of images and data on the web constitutes virtually inexhaustible material, which renews every day. It is their millions of visitors who give the digital giants their overwhelming superiority”, the author stresses, before pointing out that AI is stupid by nature and becomes intelligent only when it processes in immense quantities of data. Furthermore, “AI’s stupidity is Europe’s biggest problem, as our continent does not possess the gigantic databases needed to educate it”.
Alexandre also argues that the accelerated development of AI will lead to the relegation of human intelligence, which is incapable of processing billions of data in seconds. To increase the intellectual capacities of humans, this will require the use of implants, such as Elon Musk’s Neuralink, and/or the increase of IQ by genetic selection, a eugenic solution that seems, at least for the moment, to be the author’s preference. He does, however, acknowledge that these solutions bring with them ethical problems. Even so, both technologies will make it possible to move to a “society of general increased intelligence”, he explains, describing it as an “ultra-egalitarian society”, once it has removed inequalities in the form of intellectual abilities.
“The new revolution is not a door to a new world: it is a door to heaven. It causes many ethical, philosophical and spiritual shocks that will shake the dynamic of politics (…). The demiurgic and Promethean tendencies of life-science engineers will increase. Life, awareness, humans will be infinitely manipulable”, Alexandre concedes, before going on to explain that “the way the media are regulated is entirely out of step with the current world, in which communication has become something that can be adjusted to every individual, with no serious institutional control (…). ‘Fake news’, manipulation and ‘filter bubbles’ need new instruments. A democratic state can keep an eye on what is happening on (French private television channel) TF1, but not on every different thing on every one of several million computer screens. The Western governments (and the European Commission) are shirking their responsibilities by asking the platforms to police themselves: this is tantamount to appointing Mark Zuckerberg and the Google bosses editors-in-chief of the world. Making the GAFAM the gatekeepers of the anti-fake news laws amounts to commissioning them to define the truth!”
“There have been concerns for several years that our society is heading for a relativism of the truth, the era of the ‘post-truth’. Now, the opposite is taking shape: generative AI can create the equivalent of the Ministry of Truth from 1984 (the George Orwell novel: Ed), which held a monopoly on true and false”, Alexandre writes, going on to highlight the role of journalists: “as neurotechnologies will completely overturn our relationship with reality, journalists, whose role is to seek the truth, are a vital link in the society of the 21st century. They will be vital players in neuro-ethics by guaranteeing reality. The citizens must be helped to differentiate between reality, virtual reality, reality filtered through Google and Facebook, ideas and memories implanted chemically or electronically. Journalists will protect us from neuro-manipulators. They will belong to the range of professions who will authenticate reality, along with neuro-ethics experts, virtual reality lawyers and the regulators of cerebral prosthetics”.
The author, who reserves harsh criticism for the conservatism of national education in France, for instance by pointing out that virtually nothing has changed in classrooms for the last century whilst everything has changed in operating theatres, predicts a new role for schools: “teaching the citizens of the future to avoid cyber-addiction, to find their way around the maze of cyber-space to protect free will”. Additionally, to fit them better for their traditional role of increasing intellectual abilities, schools should be open to experimentation in forms of individualised and adaptive education.
“Levelling out cognitive inequalities should be the obsession of the State. People often feel that the State should concentrate on its sovereign duties, such as justice and defence. These days, that is an obsolete concept”, the author argues, albeit without telling the reader who should be in charge of justice and defence: the robots of the future? But these will have to be acquired, which will lead to commercial dependence, or built, which will require massive investment in research – but also commanded and control, maintained and modernised. And will all of this not constitute sovereign duties?
“Whilst the great Yalta conference of the 21st century is in session, we are not at the negotiating table, we are on it. Europe is the main course on the menu given to the nations who will be the big winners of the knowledge economy. Like Africa was in the 19th century”, writes Alexandre. He adds: “Long the dominant civilisation whose superiority culminated in the industrial era, Europe is not looking in great shape for the new century. A wonderful, benevolent, maternal and gentle mother, it does not have the weapon of the moment: AI. In this war of a new kind, our continent is on the verge of definitive defeat (…). China’s share in global research spending has spiralled: from 2% in 1995 to 23% today, in other words more than the whole of Europe and catching up fast with the United States.
“The Chinese president has (…) announced that his country will become the number one military power thanks to AI”, the author recalls, going on to state that “increasingly autonomous robot soldiers will play a bigger and bigger role on battlefields and this will require us to rethink the art of warfare”. He adds that “in the face of Chinese imperialism, the West must psychologically rearm (…). Disarming in the era of AI would be guaranteed colonisation. We are running the risk of a ‘Munich of military AI’. In 2040, whoever will agree to send their children to be modern-day cannon fodder, to their certain deaths at the hands of AI-powered robot killers? Nobody! The naive, compassionate slogan ‘no robot killers’ is suicide (…). Europe will be a geopolitical dwarf because it is becoming a technological dwarf. Europe’s low growth will prevent investment in cyber-security and France will not be able to guarantee Europe’s cyber-security on its own against the Chinese-American duopoly on AI: it will take investments of tens of billions of euros”.
What does the future hold? Alexandre proffers three hypotheses. The first is a “great conservative leap backwards”, in the form of the “election of a technologically conservative party”. “The deeper the economic mutations (read: the robotisation of manufacturing activities and replacement of intellectual activities with AI) and the greater the upheaval on the labour market, the more successful political parties offering simplistic solutions to put everything back to how it once was will be. The generalisation of neuronal implants and, to an even greater extent, embryonic selection will unquestionably be difficult steps for society to take. There will be no shortage of bio-conservatives standing up against the transhumanisation of minds, describing this group of people as the “Amish of intelligence being marginalised. The second scenario is that of a race to intelligence, leading to eugenic and neurotechnological escalation in a world that remains “structured by religious and ethnic rivalries”. “New neuro- reinforcement technologies will be seen as indispensable means to dominate other countries or rival groups”, with a considerable potential for excesses and increases of conflict. A third scenario describes a slide into neuro-dictatorship, possibly at the hands of the digital tech giants, autocrats or single parties (Chinese social control offers a still modest illustration of this) or even, eventually, a future strong AI with self-awareness, deciding whether to enslave humans or destroy them.
Clearly, none of these scenarios makes for happy reading; furthermore, they are not necessarily mutually exclusive. It is hard to predict how long they may take to come about even though, as the author stresses, the speed of technological process is constantly accelerating and will do so even more with artificial intelligence. It is therefore urgently necessary to prepare for the knowledge society by holding the ethical and philosophical debates that are necessary to anticipate changes and create the legal frameworks necessary to regulate them. However, Europe must think seriously about what it wants for its future and do all in its power to make up some technological ground. (Olivier Jehin)
Laurent Alexandre. La guerre des intelligences à l’heure de ChatGPT (available in French only) Jean-Claude Lattès. ISBN: 978-2-7096-7255-9. 473 pages. €22,00