Lovelock, then, urges us not to worry too much: it will be in the self-interest of AI systems to be environmentally benevolent. In particular, he argues that they will want to stop the earth from overheating: like people, cyborgs would not want to ‘roast’ because that would be the end of them.
But how can he be so sure? How conscious are they now, or will be in the future? AI systems are dependent on being fed content by Large Language Models. They find patterns in the human use of language and emulate them. At this stage, at least, their ‘decision making’ is probabilistic rather than logical.
But can they ever ‘know’ whether the information they soak up is materially correct, or logically sound? The notion that they could develop a sense of responsibility, even in their own self-interest, seems implausible.
Energy-greedy
Stephen Hawking said in 2014: “We cannot quite know what will happen if a machine exceeds our own intelligence, so we can’t know if we’ll be infinitely helped by it, ignored by it and sidelined, or conceivably destroyed by it.”
The AI revolution is probably the fastest cultural development in human history. And now the time has come for ‘Generative AI’, or AGI, potentially vastly surpassing human cognitive capabilities, drawing on a huge variety of data they were trained on.
They facilitate our daily demands for video streaming, remote working, or text and image editing. They store everything we see on our TV screens and what we do on our laptops and mobile phones in some distant electronic cloud, seemingly forever.
These systems are not ‘carbon neutral’. The world’s data centres rely on a steady electricity supply, from whatever source. By 2030, data centres are expected to consume eight per cent of all US electricity.
A new data centre is now opening somewhere in the world every three days. Huge buildings with interiors stuffed full of energy-greedy computers are popping up in ever more locations.
Insatiable
What I’m writing here, right now, is stored away somewhere and it is kept there as long as there’s a steady electricity supply. Only lack of power plants and suitable transmission lines are constraining growth.
In the US, Donald Trump wants to address this problem head on: on Friday, 14 February 2025 he signed an executive order for a ‘National Energy Dominance Council’ to be tasked with assuring the vast electric power needs that will be required to win the AI race with China.
He told reporters: “We’re going to be energy dominant like nobody else, and this doesn’t even discuss all of the electricity that we’re going to be producing for all of the AI plants.”
Lovelock, who was focussed mainly on scientific aspects of AI development, may have overlooked the implications of the insatiable, ever-increasing energy demands of the Novacene.
Cooling
Whilst some data centres owners are focussed on renewable energy, others are eyeing modular nuclear reactors, or even reviving abandoned reactors such as Three Mile island.
The expenses are gigantic. Microsoft and OpenAI’s planned ‘Stargate’ data centre is estimated to cost $100 billion and to consume as much as five gigawatts of electricity – more than the capacity of the largest US nuclear plant today.
A local power utility in northern Virginia reports that planned data centres are asking for several nuclear reactors’ worth of new power. This may not have worried James Lovelock, a supporter of ‘climate friendly’ nuclear energy, but other people in the Gaia movement may not be so sure.
At present some 40 per cent of data centre energy use can go toward their cooling systems. With global warming these cooling requirements are likely to soar. Not surprisingly developers are looking for more efficient ways to keep temperatures low – seeking to locate data centres in cold places, on the ocean floor or even in space.
Memory
There is also another profound concern: What is barely being discussed is that AI is usurping the vital role of our individual organic brains to remember, and to be creative. In humans, memory slowly accumulates as children learn from their parents and then in school and university.
Thus we gradually build up knowledge which has to be relearned with every new person being born and growing up. Living memory then is then stored in documents and books that pass from generation to generation.
The digital AI enterprise, in contrast, could be called a ‘memory industry’. As chatbots usurp our brains, individual creativity becomes increasingly redundant. And as AI systems externalise memory, it is effectively industrialised.
Computers ‘learn’ by analysing accumulated data in large language models and then make predictive decisions as what might be the next in a sequence of words.
Oversight
Knowledge accumulation is thus dominated by these electronic brains which boss us around in a new kind of dependence. The Novacene is defined by this memory industry.
Geoffrey Hinton further warns that as AI continues to improve, it will eventually surpass human intelligence in decision-making. If machines reach a stage where they can write and execute their own code, they could become uncontrollable.
“They will know how to manipulate people. They will learn from Machiavelli, from politics, from human deception,” Hinton explained. At that point, he argues, humanity may be powerless to stop them.
While AI offers incredible potential, the growing concerns from experts like Hinton highlight the need for stricter regulations and oversight. As technology continues to evolve, the question remains: how much control will humans ultimately retain?
Message
In 1968, information analyst Marshall Mcluhan coined the phrase “the medium is the message” He said: “The ‘content’ of any medium blinds us to the character of the medium.” This is becoming the seemingly inescapable character of the AI world of the Novacene.
He went on to say: “A medium is not something neutral – it does something to people. It takes hold of them. It rubs them off, it massages them and bumps them around … and the general roughing up that any new society gets from a medium, especially a new medium, is what is intended in that title.”
Content moderation of online platforms is intended to try and ensure that certain standards are met by removal of inappropriate content.
But as AI companies race to be competition leaders, content moderation is largely going by the wayside. Never mind the content of messages, seemingly unencumbered ‘freedom of speech’ has become the name of the game.
We read stories about AI programmes writing a 300-page novel or a 50-minute symphony in a couple of minutes, or produce a new Old Master painting in half an hour.
Do AI programmes know what they are doing? There is no indication of any content awareness involved in this. All of us, as routine users of chatbots, contribute to their dominance in our lives, and along the way, it’s ever-growing demand for processing power and storage capacity.
Insatiable
It seems to me that Lovelock underestimated the predatory nature of the AI revolution. It’s insatiable demand for storage capacity has no equal in history.
Cropping up as vast data centres in non-unsuspecting landscapes, demanding ever more electricity supplies, they become almost like cities, but not populated by people, but by data.
By industrialising memory, the AI revolution diminishes the power we have to control our lives. It is controlled by the tiny number of entrepreneurs willing to ignore the inconvenient truth that they may not ultimately be in control of the monster they have created.
What will be the prospect when Quantum Computing goes mainstream? There seem to be no limits to growth and proliferation of AI and the associated technologies. And, questioning Lovelock, this does have significant environmental implications.
Gaia
Meanwhile back on terra firma, development of artificial intelligence is being viewed with growing concern for other reasons as well: will the sheer processing capacity of AI systems bring about a systemic conflict of interest between cyborgs and humans?
Lovelock remained optimistic: “Electronic life depends on its organic ancestry … For cyborg life to emerge it requires the services of a midwife. And Gaia fits that role. … Once the cyborgs have become established, we will no more be the masters of our creations than our much loved pet is in charge of us. Perhaps the best option is to think this way, if we want to persist in a newly formed cyber world.”
And: “If I’m right about the Gaia hypothesis and the Earth is indeed a self-regulating system, then the continued survival of our species will depend on the acceptance of Gaia by the cyborgs.
“In their own interests, they will be obliged to join us in the project to keep the earth cool. They will also realise that the available mechanism for achieving this is organic life. … Because of their own self-interest, they will be eager to maintain our species as collaborators.”
Ascendant
But Lovelock has one proviso: “The notion of allowing the evolution of adaptive computer systems on military platforms seems to me to be potentially the deadliest idea yet introduced for the replacement of human and other organic life on Earth.”
The question increasingly arises: who are we really, amplified as we are as never before by an ever greater array of technologies?
We are in the process of handing the gift of knowing on to non-human beings. The deal is that we are getting evermore rapid access to ever more information.
But let’s not kid ourselves: We are all extras in competitive games being played by a few super rich individuals who know no limits to growth.
It is a very one-sided relationship: AI systems and cyborgs are in the ascendant, and if we’re lucky we may play a small bit part ourselves.
Vital
Max Tegmark, AI analyst, remarked: “A lot of the politicians are taking it for granted that if they just get AGI first, they’re going to control it … but they don’t know the first thing about the technology, it’s just sort of going on vibes.”
I end with a quote from François Rabelais, the French Renaissance writer: “Science without conscience is the ruin of the Soul.”
Lovelock’s proposal to call the current age, successor of the Anthropocene, the Novacene is a clever idea. But how compatible is this New Age with Gaia, our living planet home?
As the inventor of important scientific instruments, Lovelock was fascinated by the breakthroughs that brought about this new age, but he did not concern himself very much with who was to control its evolution.
Can we be so sure that AI systems share a vital common purpose with the preservation of a place to live, as he suggests? There is little evidence so far that the new masters of the universe are concerned about this vital matter.
This Author
Professor Herbert Girardet is an author, filmmaker and consultant on aspects of cultural ecology. He is a cofounder of the World Future Council. He is a trustee of the Resurgence Trust, which owns and publishes The Ecologist online. You can find Girardet’s Megamorphsis series on our website.
No Comments
Leave a comment Cancel