For school leavers, the choice of career is always daunting, but today’s Leaving Cert students are facing multiple global-scale crises: the arrival of artificial intelligence (AI), climate change, war, economic turmoil and the concentration of power and wealth in ever-fewer hands. For the first time since the Second World war, higher education is not delivering the pay premium and secure employment it once did.
Because of advances in technology, the friendly receptionist is gradually being replaced by a screen, chatbot and intelligent security system. Bookkeeping can be done easily by an app that links to your bank account. The data entry clerk may be a vanishing tribe. AI can design a website, a marketing strategy and a business plan in seconds, for free or very little cost. It can generate legal documents, intricate engineering designs and write code. It will revolutionise healthcare and education and open up many new ways to access knowledge and solve intractable problems.
The technology could also eliminate half of all entry-level white-collar roles within five years, according to Dario Amodei, chief executive of Anthropic. However, many of the “green” skills that the energy transition requires will still be in demand. Jobs that require on-the-spot problem-solving skills, along with caring roles and strategic and creative thinking, are probably the most secure for the moment.
When asked on a podcast recently what careers young people should train in, the so-called godfather of AI Geoffrey Hinton answered emphatically: plumbing. Robots can’t (yet) navigate awkward spaces like attics or conduct manual tasks that require high levels of skill and dexterity. There’s probably an AI chatbot capable of diagnosing your plumbing problem, but only an actual plumber can fix it.
Ryanair’s booking practices ‘an absolute disgrace’, says District Court judge
O’Donovan denies being behind what An Post described as ‘irresponsible leak’ about company’s finances
Trump reaps $50bn tariff haul as world ‘chickens out’
Irish tech worker detained by immigration agents in US for 100 days: ‘I didn’t know when I was getting home’
AI will have social, psychological, ethical, organisational and ecological impacts that are difficult to fully predict. And that is the most benign version of what might happen. Many leading AI developers are now concerned about the potential for AI to do serious harm, with lethal autonomous weapons just one of the grim possibilities, alongside concerns about privacy, manipulation and algorithmic bias.
[ Why ‘godfather’ of AI Geoffrey Hinton quit Google to speak out about risksOpens in new window ]
One of the impacts of AI that is too little discussed is the ecological one. It has what has been described by environmental news website earth.org as having a “a staggering carbon footprint”. OpenAI researchers have estimated that since 2012, the amount of power required to train AI models, which often have billions of parameters, has doubled every 3.4 months.
A ChatGPT query is estimated to consume about five times more electricity than a simple web search.
The International Energy Agency estimates that the aggregate energy demand and climate impact of data centres alone will match that of Japan by 2026.
While the tech companies argue that digital technologies can lead to efficiency gains (for example, “smart buildings” and “smart grids”), these gains can be offset by increased demand for digital services and devices, leading to a net increase in energy consumption, material use and emissions. And while renewable energy sources can displace fossil fuels, there is simply not enough land, water and raw materials to meet the ever-growing demand for data centres and digital services.
In Ireland, data centres are already consuming 50 per cent of the electricity in the Dublin region and over 20 per cent nationally, requiring higher levels of fossil energy to balance the grid. The Cambridge Minderoo Centre has warned of a potential 25-fold increase in global tech sector energy use by 2040 due to AI, which will put net-zero goals in jeopardy.
At some point we will have to decide whether the profits of the tech sector’s investors and shareholders are more important than the future of humanity and this fragile planet.
By the time we start asking the right questions about AI, it will already be too late: the genie is well and truly out of the bottle. The data centre boom, with all its energy and water demands, will already have happened. Our data will have been fed into the AI systems and used to target us with bespoke advertising and even political content. AI tools have already been dropped into typical computer software and applications without any of us asking for it. If the strategy of the tech sector is to make us dependent on chatbots and AI so that it can start charging for their use, they have probably already succeeded.
Sadhbh O’Neill is a climate and environmental researcher
