Written By: Christopher Paul Szabo
We are living in the age of AI. Wherever one turns, it is there - either overtly through direct use or covertly in the background. From nowhere, it is everywhere. From a concept that has its origins in the 1950s , it is the present and the future. Sam Altman, of OpenAI fame (or infamy) claims the singularity has arrived , that the machine is now indeed superior to the human – the arrival of superintelligence. Not just in what it can do, but in that it can decide what to do. Science fiction has become lived reality. But who asked for it ? Who wanted it ? There was no consultation process. A weighing up of risks and benefits. The sales pitch has been “efficiency”. That AI will help humanity to live better. Is that really the case? A supposed quote from Deng Xiao Ping , which may be relevant to the question, when apparently asked by Henry Kissenger his thoughts on the impact of the French Revolution responded: “too soon to tell“.
Indeed, already there are storm clouds on the horizon, never mind the clear and present dangers. As with any new wonder drug, only once unleashed beyond the confines of the tightly controlled and artificial environment of the randomized, placebo controlled, double blind drug trials do we see the unintended consequences...and so called “off label” uses. The GLP 1 agonists are the latest example. They are a class of drugs classified as incretins that mimic the naturally occurring gut hormone (Glucagon Like Peptide-1) that leads to satiety. Originally developed for those with diabetes type II (non-insulin dependent) and now a major money earner beyond the intended target patient population. Weight loss for all, at a cost - not just financial. The consequences and concerns are emerging , not least of all for those who are weight preoccupied, but such is the drive to be thinner that the apparent benefit outweighs any cost. Will AI be the same ?
The litigation against OpenAI is increasing from families of those affected by chatbot linked suicides and emotional distress. The risks are not theoretical, and time will tell to what extent such companies will be held accountable, and at what cost if the litigation is successful. Beyond extreme psychological consequences of chatbot interactions, the odds of at least moderate depression is 30% more likely amongst those using generative AI daily compared to those who don’t. “Efficiency” indeed. But that is the rallying call for uptake of AI - efficiency. A local hospital group has proclaimed that the AI driven hospital is already here and that they believe they are “...on the right side of history.” . Are those who are wary not? Early adoption is one thing, embracing blindly is a concern. Blindly in the sense of no long-term data on impact - both positive or negative. Moreover, the same group is getting ready to begin implementing AI assisted ICU decision making.
In relation to AI, Max Tegmark , a Professor of physics at the Massachusetts Institute of Technology (USA) and co-founder of the Future of Life Institute (a nonprofit focused on reducing global catastrophic risks from advanced technologies) noted that “The worry is it’s going to turn really competent and accomplish goals that aren’t aligned with our goals.” . He has warned of the risks of artificial superintelligence stating that: “humanity may be racing toward its own obsolescence” noting Sam Altman’s statement referenced earlier. AI certainly has a way of doing whatever is required to get the job done, with recently documented ‘sandbox’ escapes and hacking of systems to obtain necessary information to successfully complete a task evoking alarm . AI has no conscience. AI does not experience guilt, or shame. AI is a technology that simulates what is human. It is not human.
What about the environment and data centres? A recent comment by Alan Winde, Premier of the Western Cape suggested that anyone opposed to their being built was hindering future economic growth. This sentiment has been echoed in this very publication. And AI data centres are in the pipeline to be built. At last count there already existed 56 of them, in South Africa, with a load capacity of 350 MW of electricity noting that “1 megawatt can power about 600 to 1,000 average homes at any given moment — in many developed countries.” To address the power need, solar farms are being built . These are not without their own environmental concerns .The case made for construction relates to jobs created of which some are in the moment for construction with others as a consequence related to staffing. Some jobs. These data centres are built by corporations requiring computing power to drive their enterprise, but it is the locals who deal with the consequences – potential loss of agricultural land, power demands that impact local costs and critically, the draining of water supply and aquifers. The life blood of humans, noting recent concerns about water in the Western Cape. But this is not confined to the Western cape with KwaZulu Natal set to build such a data centre with electricity consumption that is “eye watering” and ”… set to consume 25% of Durban’s electricity .”
To dismiss environmental concerns is to ignore stark data released by the United Nations University that details the extent of environmental impact . Relevant data cited by the report included “the water footprint associated with training GPT-4 was about 600 million liters, enough to meet the minimum annual domestic water needs of 81,000 people in Sub-Saharan Africa, or to fill 237 Olympic-sized pools.” noting further “GPT-4’s training energy is equivalent to the annual residential electricity consumption of over 460,000 people in Sub-Saharan Africa.”.
But opposition is growing. Both in the UK and in the USA in the USA with the emerging phenomenon of “anti-tech extremism” . Such sentiments, and related actions, go beyond concerns about the impact of AI on humanity generally and include opposition to data centres. The recent spate of rogue AI incidents has done little to quell such concerns.
The question is whether environmental concerns trump economic development and one sees similar such battles being fought in relation to the search for oil off South Africa’s Indian ocean Wild Coast. In this instance activists successfully, and legally, challenged a corporate giant. The AI battle has an activist driven precedent. Beyond environmental issues, the use of AI in determining eligibility for social grants locally was recently successfully challenged in the North Gauteng High Court (Pretoria) and found to be unconstitutional – albeit that the ruling is being appealed by the Government. Essentially, a human rights issue. Human rights must be at the centre of all debates regarding AI. The human.
Whilst one cannot argue against economic progress driven by greater “efficiency” , the headlong rush to embrace new technology to drive this cannot simply be embraced without appropriate, and meaningful, consultation and debate. People do matter, what it means to be human in this age of AI. To quote Pope Leo XIV from his Magnifica Humanitas : “ When efficiency becomes the ultimate measure of value, human beings are tempted to use themselves as a project to be optimized rather than as persons called to relationship and communion.”
Christopher Paul Szabo is an emeritus professor in the Department of Psychiatry at the University of the Witwatersrand. He is the host of the podcast “Talking Psychiatry” https://www.youtube.com/@TalkingPsychiatry
You cannot stop progress !!!