Future of Mobility #29: Should we expect the unexpected with AI?

For a lifelong and enthusiastic science-fiction reader, it is a challenge to acknowledge to what extent this literature has influenced the tech titan billionaires who aspire to bring to reality its imaginary worlds. Their prospective scenarios are overflowed with generative AI, avatars, virtual reality, deep learning, neural networks, cities in outer space, humanoid assistants, cryptocurrencies, blockchains, pervasive drones, robotaxis and other autonomous vehicles that would require innumerous amount of energy to run. This is the Achilles’ heel of their visions.

One of the most famous advocates for these so-called cutting-edge technologies, Jeff Bezos, claimed during a conference in 2019, that “earth is no longer big, humanity is big”. This statement carries with it a touch of arrogance, a hint of hubris, which leads him astray. The energy required by a human to create, dream, love or make resilient decisions is almost negligible. On the contrary, any high-tech solution will consume much more power to complete similar operations. As many of us, I am fascinated by these impressive artificial models and groundbreaking innovations, but I cannot stop thinking of this primordial discrepancy on energy needs.

On top of his disruptive technologist image, Bezos gained more reputation for being a libertarian. The promoters of this movement consider that individualism is the quintessence of progress for a civilization. Without any doubt, this aspiration fits very well with a community governed by artificial intelligence where everyone is out for themselves and may the best one win. Bezos is legitimate to encourage libertarianism with AI, as these generative models work with a strong dose of heartlessness to say the least. It is a perfect match for a deregulated society, but I don’t buy it. I believe in a stronger force coming from the mutual dependence between living beings. We generate emotions that have tremendous influences, very often beyond our rational understanding. I am amazed by the potential of our incommensurable love or an intimate prayer which can spread peace and serenity without any artefact. This is what makes me wish that the universal human passion for good and mercy would always surpass any form of artificial intelligence.

These thoughts kept going round and round in my mind and everything I experienced during the last few weeks brought me back to them. During this investigation, I enjoyed the support of several companions: the author Tomi Ungerer and his children’s books for the fantasy of human creation; the philosopher Rapha?l Enthoven for his defense of human intellect not reducible to a simple mechanical function; the historian of science Jean-Baptiste Fressoz for shedding light on the illusion of our clean energy transition; and the science-fiction writer Philip K Dick for unveiling the fragility of our reality. The outcome of this cogitation is the purpose of this new article.

1. The Fantasy of human creation: AI with Tomi Ungerer

It does not take more than a few minutes to read the famous picture book written and illustrated by Tomi Ungerer, “The Three Robbers” (1961). The story tells the life of three bad guys who dispossess honest people stockpiling their loots in a cave perched somewhere in a mountain. One day, on a desolated road, they come face to face with a little orphan girl alone in a carriage. Her name is Tiffany, and this fortuitous contact will transform the life of the bandits. With her kindness and candor, she convinces them to use their stolen treasure to buy a castle to host and take care of all the abandoned children in the region. Years after years, the community grows with new residents coming in support of this initiative and building houses around the castle. At the end of the story, to celebrate these three heroes, who turned out to be good guys, the villagers erect three tall towers in their honor.

To draft this adventure totally unconventional where the boundaries disappear between the bad and the good, the foolish and the wise, the vicious and the innocent, Tomi Ungerer used his gift for creation by drawing impressive images with just a few lines and basic colors, and by writing very short and direct sentences on each page of this little book. My assumption is that a virtual assistant like OpenAI ChatGPT or Google Gemini would not be able to invent a similar story as “The Three Robbers” with such an economy of resources and without any reference model. Their algorithms would not give the chance to a mean person to become a gentle one, and even less to confer authority to an innocent child like Tiffany. It is an incredible story which can only been conceived by a human mind turned to gratitude.

This very little example is just a tiny illustration of the indelible difference that I see between the human creation and the outcome of a generative AI tool. However, if we allow too much influence and domination to AI, this gap might reduce. It will diminish our sense of wonder to the beauty of the planet, make us lose sight of nature’s gifts, and undermine our ability to bring dignity in corrupted minds like Tiffany successfully did with her snatchers. Even worse, I am concerned that it would amplify the oppositions, exacerbate the differences, boost our antagonisms, and simplify our vision of the world to a disembodied functional role. This is the concern that I wish to share throughout this article.

Last week, seized by a regular delirium of tidying up, I went up to the attic with the intention of sorting through documents. I came across a few notebooks that sent me back to the time when my children were very young. In those days, we used to invent new words, writing them down and adding a specific definition and a unique drawing. Squatting in the middle of these scattered files, I smiled inside whilst reading some of the expressions we compiled in a joyous mess. Out of all the nonsensical ideas that bloomed out of our minds, we had two favorite themes: the names of phobias and the curses from Captain Haddock.

Scientists are very imaginative and give hundreds of designations for multiple kinds of phobias. Many of them are serious and really affect those who suffer from it. For instance, Amaxophobia, is the fear of staying in a car; Catagelophobia is the fear of being ridiculed; or Didaskaleinophobia is the fear of going to school (here is a full list, but you can also look at Wikipedia with quite an extensive compilation). Others are more intriguing and touching the frontier between reality and fantasy. I like in particular the following terms: Arachibutyrophobia which is the fear of having some peanut butter sticked to the roof of your mouth; Nanopabulophobia, the fear of coming across a dwarf with a wheelbarrow; Spheniscilogophobia, the fear to say the word “penguin”; or Anatidaephobia, the fear of being watched by a duck. This last one made me burst out laughing: the Luposlipaphobia, the fear of being pursued by timber wolves around a kitchen table while wearing socks on a newly waxed floor. It was imagined by Gary Larson in his comic “The Far Side” and he drew a very expressive image of this phobia which is hilarious and scary at the same time. ?With the intention to challenge generative AI, I asked Gemini to give me an image of this phobia. The result was so dreadful and disastrous that I don’t want to insert it here by respect for your sensibility. It was another typical AI image similar to many others that we see everywhere nowadays: always exaggerated and sophisticated, overcharged with details badly adjusted between each other, a cold and unexpressive design, aggressive features, a complete lack of fun and joy, just a mechanical and bland transcription of a prerequisite... On the contrary, Gary Larson is like Tomi Ungerer, he used his tremendous talent with just a pencil to create a unique and exhilarating masterpiece.

In our family file of phobias, I also found a few disorders that we imagined together with my children. Of course, they are my favorite and you will not find them in any dictionary. For instance, the Squidofootophobia, the fear of having a calamari tickling your feet while you are swimming in the ocean, the Speakyteddyphobia, the fear that your stuffed animals would start talking in your bed while you are sleeping, the Tweetweetdroponosophobia, the fear of receiving bird droppings on the tip of your nose, or the Jobopyjamadayphobia, the fear of being at your desk in the office dressed in your relax night clothes with your sleepers firmly in your feet.

We did the same for the curses from a famous cartoon character, the Captain Haddock. As you surely know, he is the extravagant companion of Tintin unable to hold back from uttering swear words all over the place. There are more than four hundred curses in the French language and the creative English translators invented a little more than three hundred and fifty ones (see the full list in English here). The effort of the authors to create unique curses in different languages is an amazing achievement. Among my favorite ones, I would pick up these ones: Dry-dock sailors! Interplanetary pirats! Anthropithecus! Rotten sand-hoppers! Vermicellis! Miserable molecule of mildew!

I was so addicted to these expressions that I was using them to wake up my children, to call them for dinner, to ask them to hurry up or to participate to any gathering activities where I needed their attention. As for the phobias, we invented our own Captain Haddock’s curses like these ones: Tribe of tense tendon! Low-floor scallywag! Highway velocipede! Stack of onion skins! Pickle of the mountains! Feather sweeper! Of course, they are not reported in the official compilations.

These are just small samples of the fantasy of human creation. This is the treasure of being able to make fun of ourselves with fondness and a touch of irreverence. If a generative AI could invent similar phobias or barnacles, I bet that they would not be so funny or thrilling. By permanently recycling same sources of data, even if it includes the whole library of humanity since the origin, AI will impoverish our imagination and alter our open-mindedness. In the contrary, human creativity enriches our spirit with these literary wordplays and these inconceivable and endearing fictitious characters. I am confident that we will always remain sensitive to the impertinence, the cultural deepness and the imperfection of our human inspiration. Something that any generative AI will hardly be able to compete with.?

In another circumstance, I challenged the originality of AI production with a real-life example. To set the context, I am preparing a future article that will be dedicated to regenerative land initiatives. For that purpose, I am watching a lot of documentaries describing projects across many countries in the seven continents. One of them is related to a project of rainforest restoration in Guapi A?u, close to Rio de Janeiro. The owner of the land, Nicholas Locke, explains how he handled the restoration process and how species work in symbiosis to rebuild the original ecosystem, an interesting example of biodiversity regeneration. There is a fascinating very short sequence in the video (see below) that stroked my eyes: just after the four first minutes, you can see a very funny scene where three little birds seem to play leapfrog, just for a few seconds. A charming moment of wildlife. I immediately thought about what an AI chatbot would make out of it. With a friend, we decided to test ChatGPT and make a request to the system asking to find out any information about birds acting the same way or describing similar situations in the real world. Same as the previous query with the Gary Larson’s joke, the result was so shameful and pathetic that I don’t dare reporting it here. Feel free to test this out by yourselves to form your own opinion.

It seems that the word has spread out because, a few days ago, I read a post showing the result of an internet search (with Reddit) for another bird, the “baby peacock”. It appeared that more than half of the images were not real and coming from AI generated sources. A funny commenter suggested to modify the query with the name “peachick”, as if a simple synonym could save us from losing ground with the real world.

2. The Defense of human intellect: AI with Rapha?l Enthoven

Recently, I was offered the opportunity to talk about my views on Sustainability and Climate change to a group of students from an engineering university. I prepared this seminar with more than a hundred slides drafted in the old-fashioned way, meaning by myself. One of the topics was a review of the seventeen Sustainable Development Goals. At the end of the day, we proposed the students to gather into sub-groups, select one of the Goals and develop a pitch about what they would do to improve their engagement towards their chosen objective. One week later, we organized a follow up session to review the results of their investigations. Several of them used an AI tool to compose their slides and proposals. Only a couple of them spent time to reflect about their task and prepare the presentation material on their own. The difference was crystal clear for the audience: AI enabled the lazy teams to design slides quite quick and apparently nice at a first glance. However, their presentations turned out to be standard, with a format quite similar between each one of them, a content which was boring and repetitive, full of sentences rehashed by an AI chatbot which lacked substance that could mean anything and everything, and graphs that were simply copy/pasted from google without being fully relevant. Reversely, the two teams who decided to dedicate time for their speeches and build their presentations with minimum assistance were far more appealing, sincere and convincing. We noticed a few minor mistakes on their slides, but it oddly brought more personality and appreciation to their work. They were the two final pitches that were acclaimed by the jury!

This experience comforted me in my assumption that AI cannot be a replacement for our human intelligence and imagination. Now, I would like to take two examples from the literary world that I particularly cherish. The first one is about a group of authors who created, around twenty years ago, a fictitious philosopher named Jean-Baptiste Botul who was supposed to be born at the end of the 19th century (see the Wikipedia report here). To make his personality as real as possible, the team of farceurs published books, critics in the press, correspondence with other existing philosophers, references at the bottom of other scientific articles, etc. They even officially established a school of thinking called “The Botulism” in his memory. The trickery was so perfect that other recognized philosophers fell in the trap and took his existence for granted.? One of these unfortunate chumps was a well-recognized French philosopher, Bernard-Henri Levy (BHL), who took Botul’s work very seriously and mentioned him in one of his publications, which made him the laughingstock amongst the entire French literary community. This affair became so ridiculous that it went around the world as you can see in this example from the Los Angeles Times (here). Later, BHL admitted his mistake and escaped from this trap with an elegant repentance and a cordial esteem to the group of merry guys. I like this literary joke for two reasons: ?this parody is so well done that it is a pleasure to read the fake books from Jean-Baptiste Botul as if they were written by an existing and recognized philosopher. I particularly recommend his “Correspondence to myself” and the amusing “Sexual life of Immanuel Kant”. Secondly, I appreciate the humble reaction of BHL who demonstrated with honor that we can always escape from any situation, even the most ridicule ones, with elegance and humility. I will add a third reason: in no way, such an enjoyable and elaborated pleasantry could never have come out of an AI device as sophisticated as it could be. The creation of the philosopher Botul is the most admirable deep fake invented at a time when this designation did not yet exist.

The other example comes from another French intellectual, Rapha?l Enthoven. A bit like BHL, Rapha?l Enthoven is also perceived as a grandiloquent philosopher. He used to talk with an emphasis and cultivated language. I would paint the experience of his listeners in three phases: after a quite eloquent beginning, the talk enters in a tortuous body where Enthoven seems to be lost in the twists and turns of his thoughts with everlasting sentences, taking the audience with him in his wanderings, until he jumps to a conclusion adorned with magnificent quotes, while his followers remained stuck somewhere in the middle of the lecture, finding their way through his stunning torrent of indecipherable words.

Recently, he published a book whose ambition was to demonstrate the forever weakness of the generative AI tools compared to the human intellect. As you can anticipate, this provocative thesis generated a buzz within the tech-savvy community. One of the most captivating debates happened with another doctor in philosophy, Thibaut Giraud, who turns out to be also an expert in AI. Interestingly, while I was in favor of Giraud’s stance at the beginning of the dispute, I changed my mind during the controversy and ended up supporting Enthoven’s principles. Let me explain how.

It must be said at the outset that Thibaut Giraud is an adept of the functionalism theory. In accordance with this philosophy (see Stanford definition here), he considers our mental states to purely be a functional activity of our brain, which means that if we could build a synthetic system that could realize the same functions as our mind, this structure would accomplish the same states of consciousness as ourselves. In other words, if we replace all of our organic neurons with silicon cells and apply appropriate electrical impulses, the reactions and emotions of this artificial mechanism should be absolutely comparable to our human brain. Without the slightest hesitation, Giraud speculates that an AI system could think and feel similar sensations as a human. These mechanistic principles were largely rejected by Rapha?l Enthoven who defended empirical forces that cannot be reduced to materialism and rationality. This was the biggest stumbling block of their discussions. They couldn’t disagree more.

It seems that we are facing two irreconcilable thesis: from one side, the advent of a conscient artificial intelligence is a dream for functionalists and libertarians, and from the other side, the persistence of a superior human spirituality is a wish for humanists and empiricists. I tend to adhere to the second hypothesis. I intimately believe that there is a force in life that cannot be reduced to a function. We can feel it in the love we have in ours hearts, in our sentiment of the impermanence of time, in our compassion for the victims, in the subtleness and accuracy in our intuitions, in our existential fear of death or in the power of our prayers. Enthoven relies on a couple of philosophical ideas to strengthen his position: the “I-don’t-know-what” defined by the philosopher Vladimir Jankélévitch and the “Clinamen”, inclination proposed by the Roman poet Lucretius. These concepts are here to tell us that there is something inexplicable in human intellect that is irreducible to a function or a substance, a freedom that escapes the inescapable chain of causes. This is this mystery of life that cannot be grasped by any artificial intelligence.

The TV series “Black Mirror” represents what I like in science-fiction motion pictures. So, I was pleased to hear Rapha?l Enthoven mentioning one of its episodes during the debate with Thibaut Giraud. This is the one called “Be right back” (Season 2 ep 1). It tells the tragic story of a young woman, Martha, who loses her companion, Ash, in a car accident. Unable to bear the grief, she is encouraged by her best friend to test an AI application that will reproduce perfectly the personality of the passed-away boyfriend thanks to the computation of all the historic data of his life. The first experience by voice is so successful that she accepts an upgraded service offering the physical presence of an avatar almost identical to the body of Ash. All seems going well until the grain of sand towards the end of the movie. The fake Ash makes a very little gesture which does not match the genuine one and all the enchantment crumbles. Martha breaks down in tears and screams: “he would have worked out what was going on. This wouldn't have ever happened but if it had, he would have worked it out. You're just a few ripples of you. There's no history to you. You're just a performance of stuff that he performed without thinking, and it's not enough.” Enthoven was clever in using this popular example to illustrate that there is a singularity in each person that an intelligent machine will never replace in its totality as it will miss this inexplicable thing as we said earlier. This nice metaphor did not convince Giraud who remained inflexible and insensitive to it.

Although they were not mentioned in the debate, I would like to add below four reflections from Rapha?l Enthoven towards AI that I found really stimulating:

1. It is common to say that astonishment is the root of philosophy. However, the secret is not to be astonished by what is astonishing, because everyone and even the most stubborn guy will be surprised by it. This is the reason why generative AI productions are attractive, because they play with our emotions in offering innumerous images and immediate complete documents which are surprising for anyone. On the contrary, the principle of philosophy is to be astonished by what is common, ordinary, normal. Enthoven takes examples like the wonderment of the return of spring, a beautiful sunset in a clear sky, the repetition of daily and simple interactions with your loved ones... The essence of astonishment is in the everyday routine of life.
2. You could read all the books published on Earth since the beginning of humanity, like ChatGPT can do in a few seconds, and yet be unable to understand anything. “You don’t create a healthy mind just by compiling data.”? The wisdom is not in a printed text, but in your spirit, in the way you interpret and assimilate the experiences of your life, reading books being one of them. Wisdom does not require philosophical thesis or literary work to flourish. It can reside in the soul of an illiterate person who knows how to enjoy the most precious treasure which is the journey of his days.
3. Our anxiety of seeing AI dominate humans is as old as our history. The mythology is full of stories where humans bring creatures to life like the sculptor Pygmalion and his ivory statue, the rabbi of Prague and his Golem made out of clay, the scientist Victor Frankenstein and his monster or the carpenter Geppetto and his wooden boy Pinocchio. This worry tells us more about the desire for humans to be treated like a God than anything else. “We should recognize that there is more pride than humility in this concern”.
4. I really enjoyed listening Rapha?l Enthoven when he commented this last consideration: for him, AI cannot understand the difference between ingenuousness and candor (like in the story with Tiffany and the three robbers). “Ingenuousness is when you believe that a streetlamp is a crescent moon, and candor is when you know that it is a streetlamp, but you intentionally decide to take it for a crescent moon.” AI will fail to figure out what is the value to persist deliberately in an illusion encouraged by a sincere intention.

I love these kinds of invigorating concepts. Of course, they will not help to close the debate between Giraud and Enthoven, but they are so refreshing for our spirit. They are powerful as a therapeutic treatment.

To nourish the debate, I would recommend reading the book of the research professor Kate Crawford, “Atlas of AI”, where she debunks two misconceptions that our two philosophers did not manage to get out of it: “The first myth is that nonhuman systems (be it computers or horses) are analogues for human minds. This perspective assumes that with sufficient training, or enough resources, humanlike intelligence can be created from scratch, without addressing the fundamental ways in which humans are embodied, relational, and set within wider ecologies. The second myth is that intelligence is something that exists independently, as though it were natural and distinct from social, cultural, historical, and political forces.” Her provoking assumption is that AI is fundamentally political. AI is a mix of technical and social practices devoted to “the global interconnected systems of extraction and power, not the technocratic imaginaries of artificiality, abstraction, and automation.” Let me add the following paragraph that would give us an original perspective on the influence of these advanced technologies: “AI is neither artificial nor intelligent. Rather, artificial intelligence is both embodied and material, made from natural resources, fuel, human labor, infrastructures, logistics, histories, and classifications. AI systems are not autonomous, rational, or able to discern anything without extensive, computationally intensive training with large datasets or predefined rules and rewards. In fact, artificial intelligence as we know it depends entirely on a much wider set of political and social structures. And due to the capital required to build AI at scale and the ways of seeing that it optimizes AI systems are ultimately designed to serve existing dominant interests. In this sense, artificial intelligence is a registry of power.”

3. The Illusion of clean energy transition: AI with Jean-Baptiste Fressoz

During his conference in Washington DC on May 9, 2019 (“Going to Space to Benefit Earth” - see video below), Jeff Bezos showed artificial pictures of cities in the space. By watching this recording five years later, I had a weird feeling of disillusionment and a clear disconnect with what should have procured these photos to the audience on that day. The power of attraction of generated images shrunk with the time. They lose their fascination soon after their first apparition. The reason is because they don’t have any emotional substance, they don’t have this “inexplicable thing” commented by Enthoven which makes human productions totally unique. The simple village with its three tall towers drawn by Tomi Ungerer more than sixty years ago (in 1960) keeps its freshness and attractiveness up until today and will continue forever to transmit an intangible feeling of sympathy.

At the same conference, before presenting his futuristic urban places, Bezos exposed his views on the becoming of humanity. He recognized that our “very fundamental long-range problem is that we will run out of energy on earth”. ?A bit further in the conference, to the question “what won’t change in the next ten years?”, he predicted that “customers will continue to want low prices, fast deliveries and big selection, so all the energy we put to meet their requirements will continue to pay dividends”. There is a mismatch with what he said earlier because, to meet these never-ending customers’ expectations, it will require more energy while, at the same time, he is worried about its scarcity. This self-centered and obsessive requirement from customers is a given in Bezos’ plans, this cannot be changed, and the society will have to adapt to it to be successful. This is a very important and bothering statement, because it implies, probably rightly, that our societies are not ready to accept any scenario with energy sufficiency, where we would collectively agree to change our behaviours of consumption in order to reduce CO2 emissions.

In 2020, Jeff Bezos created the “Earth Fund” to accelerate the energy transitions in developed countries with the hope that it will fight climate change and solve the risk of energy scarcity. I don’t want to appear too skeptical, but I found this ambition rather deceptive. My intuition was strengthened by the reading of “No Transition”, the latest book published by the historian of science, Jean-Baptiste Fressoz. ?In this resounding thesis, the author demystifies the concept of energy transition in which we are embarked. He hits all along his pamphlet on this confusing and misleading illusory nail. He takes a dozen of examples to demonstrate that this dream of energy adaptation by phases (from the age of wood to the age of coal, then to oil, and eventually the age of silicium) never happened in history. In fact, this is the opposite. The evolution of energy and raw materials is exclusively cumulative. We accumulate new sources of energy on top of the previous ones, and all of them are working together in symbiosis, while their consumption volume keep increasing relentlessly.

He justifies his argumentation with many examples: the utilization of wood was essential to build the mines, the oil rigs, the railway tracks, the barrels, and all the installations that were required to extract, produce, transport and consume fossil energies. Its consumption has never been so high since the beginning of the coal industry. The same synergy occurred between fossil energies: oil is intertwined with coal because, all along its supply chain, it relies on steel machines (wells, vessels, tank trucks, pipelines, refineries, cars...) and cement for the mobility infrastructure (roads, parking, buildings...). Reversely, coal is dependent on oil for its extraction and a monumental example is the strip mining to release lignite. In the same way, the development of renewable energies like solar panels and wind turbines is intrinsically associated with the preceding primary energies and materials.

Historically, the popularity of the energy transition legend is credited, by Fressoz, to the president of United States, Jimmy Carter. On 1977 April 18th, in a dramatic allocution to the nation (see video below), Carter predicted that by the 1980s the world would fall in crisis due to a higher demand on oil than it can produce. He took profit of this wishy-washy prophecy to spread the word about the successive ages of energy: “Twice in the last several hundred years, there has been a transition in the way people use energy. The first was about 200 years ago, when we changed away from wood, which had provided about 90 percent of all fuel,? to coal, which was much more efficient. This change became the basis of the Industrial Revolution. The second change took place in this century, with the growing use of oil and natural gas. They were more convenient and cheaper than coal, and the supply seemed to be almost without limit. They made possible the age of automobile and airplane travel. Nearly everyone who is alive today grew up during this period, and we have never known anything different. Because we are now running out of gas and oil, we must prepare quickly for a third change: to strict conservation and to the renewed use of coal and to permanent renewable energy sources like solar power.”

Fressoz explains that Jimmy Carter was misled by an Italian physicist, Cesare Marchetti, who had produced a report a few months prior called the “National Energy Plan”, which contained a graph describing the historic evolution of energy consumption in relative value. I highlight these two words because they are the clue of the thesis. For the purpose of my article, I preferred to use the updated data from ourworldindata.org instead of the old figures from Marchetti. The first graph below represents the evolution of primary energy consumptions relatively between each other (in %). This visualization is perfect to suggest a progressive substitution from one source of energy to another. ?From this perspective, the transition seems obvious from wood to coal, then from coal to oil and natural gas. This is a similar information that Carter used to make his address to the nation. Now, if you look at the second graph, which provides exactly the same information, but this time in amou, you will notice that there is no transition at all and the consumption of all the primary resources is increasing since the beginning of our industrial revolution.

Now, let’s come back to Jeff Bezos and his foundation to accelerate the transition to renewable energies. After reading Fressoz, I am afraid that this philanthropic initiative is hiding an intention less honorable: the large investment on solar and wind technologies by Amazon, Google, Meta or Apple has little to do with global emissions reduction but is rather planned to support the boom in generative AI and the enormous amounts of energy required to operate the future data centers all around the planet. Not to mention their plans to revive nuclear power plants, which is a stupefying alliance between two polarized technologies. Recently, this trend made the headlines of several investment banks (Goldman Sachs, Morgan Stanley…) and media ( The Guardian). Among other data, they explained that Generative AI is 10 times more energy-intensive than a traditional search engine. I looked for the source of this information and came across this exciting research completed by three scientists, Alexandra Sacha Luccioni, Emma Strubell and Yacine Jernite (here). I discovered that It is even worse for image generation which stands at the highest end of the spectrum, consuming several hundred times more energy than a single task.

In the same vein, I cannot ignore either the impact of autonomous mobility which is a topic that I like very much. Another researcher, Soumya Sudhakar, from MIT completed a remarkable study revealing the extremely large amount of energy that will be needed to move these cars autonomously (see video below). A driver is a nest of mistakes but he can adjust his behavior quickly, even if it leads to another stupid decision, all of that while using limited energy. On the other hand, an unmanned vehicle will have to compute millions of scenarios in a fraction of second to make the most appropriate choice, which would require to duplicate the infrastructure for data centers globally. In the second part of her presentation, Sudhakar proposes some remedies to the voracity of this technology. I really like her solution to make robots not “think too hard”, with innovations in energy-efficient hardware and algorithm designs, which could mimic the work of a human brain. However, even a fool can avoid a collision just through pure instinctive survival reaction, while it will remain a mysterious attitude for an autonomous vehicle downgraded to a foolish digital configuration.

Soumya Sudhakar's main conclusion is to recommend benefitting from the expansion of renewable energies to provide incremental power capacities. Without realizing it, she confirms the warning raised by Jean-Baptiste Fressoz: Solar panels and wind turbines might be rerouted to enable additional services for new technologies like generative AI or autonomous mobility instead of decarbonizing existing activities.

The other day, I was stunned to listen to a scientist falling in admiration with some, allow me to say, bullshit developments of machine learning models. For example, you can take a photo of the interior of your fridge and the AI will compute the ingredients to propose several detailed recipes. The geek was almost in ecstasy when the system suggested to finish a spicy sauce jar that was forgotten in the back of one layer of the fridge. Another feature made him laugh in tears as if he was overcome by a true veneration: to make a meal decision in a restaurant, you take a photo of the full menu and add instructions about the budget limit for each guest. Then, the AI will propose several options of complete menus including beverages with calories and detailed prices including tax and tap.

Isn’t such a technocratic delirium ridiculous? Are we going to waste our incremental clean energy resources to generate artificial needs until the very last corner of our human activities? Are we not delegating to AI our communication, our interactions and our creativity which are essential parts of our humanity? This is where morality shatters and Bezos’ prophecy becomes unquestionably fulfilled. I am worried that these extravagant services will be largely praised by customers who consistently expect more value at a lower cost and fall in awe by a spice of tech innovation, without thinking for a moment about the impact of these absurd requirements. As many other insignificant acts (like throwing away a plastic bottle or a cigarette butt), we hardly imagine the disastrous consequences when they are reproduced at scale on our environment and our health.

A few years ago, when I was living in India, we launched a project to provide free energy to isolated populations in remote locations not connected to the electrical grid. Our proposal was to install a solar panel and a second-hand battery from electric cars. The beneficiaries would have enjoyed a minimum comfort with air conditioning and fridge, which are vital during the hottest periods of the year. This project of decentralized renewable energy networks to local communities did not go through unfortunately. But in a society dominated and mesmerized by generative AI productions, I am afraid that such a project would become a minor priority in future decarbonization strategies.

The IPCC Working Group III on the mitigations of climate change published a chapter (the 17th) dedicated to justice and equity during the transition to a low-carbon development. There are good intentions and recommendations but nothing which could address this concern regarding the monopolization of clean energy capacities by the new technologies combined with a lack of willingness from consumers to adopt sufficiency solutions. I am a bit ashamed to say it but I find the tone and style of these reports rather boring. There are so many conventional sentences, and the structure is so academic that, if it weren't for the names of the fifteen authors mentioned on the front page, one would believe that it was written by an AI chatbot. I understand that such a publication has to be neutral with standard practices, which is a normal way to publish such kind of material, but a pinch of enthusiasm would be more than welcome. A small light came out on page 43, when I read about a project in Tanzania handled by a non-profit organization to deploy small-scale solar power to impoverish communities which reminded me our project in India. It seems that this Tanzanian project was successful. Kudos to them!

4. The Fragility of our reality: AI with Philip K Dick

In his biography of Elon Musk, Ashlee Vance witnessed the following: “The Silicon Valley’s techno-utopian club tends to be a mix of Ayn Rand devotees and engineer absolutists who see their hyper logical worldviews as the Answer for everyone. If we’d just get out of their way, they’d fix all our problems. One day, soon enough, we’ll be able to download our brains to a computer, relax, and let their algorithms take care of everything. Much of their ambition proves inspiring and their works helpful. But the techno-utopians do get tiresome with their platitudes and their ability to prattle on for hours without saying much of substance. More disconcerting is their underlying message that humans are flawed and our humanity is an annoying burden that needs to be dealt with in due course.”

When they were young, Elon Musk and Jeff Bezos were passionate about science-fiction literature. Both enjoyed the reading of the Isaac Asimov’s Foundation series, “The Moon is a Harsh Mistress” by Robert Heinlein, “The Lord of the Rings” and “Dune”. The boss of Tesla and Space X even said in many instances that the most influential book of his life was “The Hitchhiker’s Guide to the Galaxy” by Douglas Adams. I shuddered hearing these comments as I very much like the novels they mentioned and was wondering how on earth it is possible that these masterpieces influenced these businessmen in such a way that they believe so ardently in building artificial words and forecasting no more future for our human destiny than one outside in outer space. I read the same books when I was young and I deduced the opposite from them. They gave me a strong feeling that our planet is unique in the galaxy, that it is nothing more than this tiny oasis of life lost in the universe, that she is fragile and we have to take care of her as much as we can because she is the only spaceship viable for our humanity.

I was also surprised that Bezos and Musk did not mention another iconic author of science-fiction in their top list, Philip K Dick, and particularly his novel called “Ubik”. It is likely his most famous book and, on many instances, I think that it outlines the world that the two billionaires would like to build for the world population. There are colonies on Luna and Mars, you can go to the moon in just one hour instead of three days, the interstellar flights are becoming workable, you can eat Martian grubworms or braised fillet of Martian mole cricket, you can predict the future, and after your death you fall in a temporary semi-life where you can still communicate with your living loved ones. But the most striking connection between this novel and the fantasy of the two magnates is at a higher existential magnitude. At a level where we get confused with our reality. In “Ubik”, the time escapes from its linearity, it amplifies in different directions, the protagonists lose control of themselves and their environment, they don’t know if they are alive, dead or if they dream, the materiality is reversed to earlier forms, the time periods and segments get scrambled, and eventually the lector of the novel is drawn into this incredible society twisted by different phases of reality. A unique and exceptional reading adventure.

If we are not careful, this might be the future that these tech titan sorcerer’s apprentices could deliver to our real society. With the advent of generative AI, deepfake videos, audios and manuscripts, multiple avatars, replicas of passed away people, metaverse immersive experiences, digital assistance for anything and others virtual realities, we might end up living in the same distorted world as the characters of “Ubik”, not being able to discern what is true, real, or tangible, and losing our perception of time and space. To defend this worrying prospect, I came across a research from Canyu Chen and Kai Shu (here) which reveals that we are easily deceived by fake news created by AI: “Generated misinformation can be harder to detect for humans and detectors compared to human-written misinformation with the same semantics.”

In "Ubik", there are two characters that are particularly interesting to mention. The first one, Joe Chip who seems to have a preordained name with the prominent place taken by microprocessors in our contemporary society. I don’t know if it was fortuitous or intended by the author. The second one, Jory, is the most dangerous and dominant character of this desperate world. He survives by soaking up the energy of the semi-dead bodies. He considers that the world is nothing more than a pure product of his mind, there is no existence of any sort outside of him: the people, the things, the events are features brought to his pseudo-world by his mental capacity. Everything will vanish as soon as he goes away: “as soon as you leave a place, it passes out of existence”. The entire world is contained in his brain. What would happen if Jory became an exemplary model of identification for many of us lost in the artefacts of our digital society?

I don’t think any artificial intelligence could invent a novel to the same level of complexity as “Ubik” because it embeds in its narration several emotional and fundamental questions about our future of humanity that can only be envisioned by a human mind. Philip K Dick has the creative capacity as Tomi Ungerer with the ability to immerse us in a totally unexpected adventure.

As a matter of fact, Tomi Ungerer is not only famous for his artistic creativity. At a minor level, he is also considered as the owner of the following motto: “Expect the Unexpected”. It is true that this slogan applies nicely to his production which is often satirical and provocative, even in his children’s books like the one with Tiffany and the burglars. I slightly modified his maxim in the title of this article to emphasize the paradox of AI in our society. We are all fascinated by the promise of these generative models to bring us unexpected achievements and sensations. But we are wrong, and I like the comment from Rapha?l Enthoven about the wonder of ordinary things. I am like a rebellious little hobbit who refuses to apply all these AI functions that the media platforms are aggressively forcing us to use. I always feel happy with myself when I ostentatiously reject their proposed words and sentences, ignore their rules to write better posts and disregard all their disincarnate AI images. So, should we expect the unexpected with AI? After what I have shared above, my answer would be certainly not.

A dreadful dystopia could be a society dominated by generative AI and neural networks which would produce uncontrollable fake information and build virtual realities in which any citizen would become a simple cell activating and reinforcing the power of the artificial organization. Philip K Dick was a precursor in 1968 by describing a similar world in his novel, “Do Androids Dream of Electric Sheep?”, from which Ridley Scott made the famous movie “Blade Runner”. Shouldn’t we also question our dreams and make sure that we are not launching AI solutions that will go against our expectations and lock us up into an unmanageable terrifying society? Since the years 1970s, the publication of the Meadows report and a few major post-apocalyptic novels (John Brunner, Philip Wylie, Jim G Ballard, John Christopher, James Blish or Harry Harrison), we were crying wolf with the climate change while not believing in it. Are we going to reproduce the same inertia with an uncontrolled deployment of AI in every corner of our lives, and later suffer the unforeseeable consequences?

There is another children’s tale from Tomi Ungerer called “The Beast of Monsieur Racine” which talks about the expectations of a retired tax collector, likely the opposite profile of a libertarian. He lives in his cottage and “content to spend his days shuffling about, watching birds and clouds, and tinkering with his little garden”. In two occasions in the story, he could make a fortune: the first time, by selling his pear tree which bears fruits with exceptional flavor that millionaires would love to acquire; and the second time, by giving up his loyal and unique companion to businessmen who dream to exhibit his strange creature to the world. However, he systematically refused, because he had something in his mind that AI will never offer: a good sense of humor, a tenderness for others, and a taste for the simple little things offered by nature.

Related articles:

• Future of Mobility #17: Once upon a time there were five young scientists...
• Future of Mobility #12: The Power of algorithms
• Future of Mobility #10: My take on the global semiconductor crisis
• Future of Mobility #5: Data Extraction & Analysis: Surveillance or Benevolence?