Last month, the AI Seoul Summit took place in South Korea, bringing government leaders together with tech companies, academics, and civil society to advance discussions on the safety of artificial intelligence.
The event concluded with 16 of the world’s most powerful tech giants promising to minimise the risks of the AI revolution.
These global corporations, including the likes of Microsoft, Meta and IBM, signed up to the newly developed "Frontier AI Safety Commitments". They will now have to publish their own safety frameworks, explaining how they will measure the risks of their AI models. Signatories must also outline the threshold at which risks will be declared “intolerable.”
But these same corporations are turning a blind eye to AI’s most urgent danger: its catastrophic impact on our planet.
Our world runs on energy — and so does AI
The ugly truth is that AI relies on hordes of existing data. That data is stored on physical servers in vast facilities buzzing with electricity.
These servers are always on but will be ruined if they overheat. So, their constant activity must be offset by enormous cooling systems that cycle through 1.1 quadrillion gallons of water each year.
Unsurprisingly, these sprawling setups require a staggering amount of energy. By 2025, they could account for up to 10% of global consumption. And while some of this power will be supplied by renewables, most will rely on natural gas.
Our glorious new technological era will be built on the same dirty fossil fuels that are destroying our planet.
This means that our glorious new technological era will be built on the same dirty fossil fuels that are destroying our planet.
And this is before we even begin to consider how AI is being used. Of all the nefarious actors at play, fossil fuel supermajors are among the most powerful.
In their unquenchable thirst for fatter profits, they race to invest in AI technologies that can yield faster and cheaper results when searching for untapped resources.
But this is only part of the picture. After all, there are plenty of opportunities to use AI for good.
Big Tech vs Big Oil
Already, experts are developing novel solutions to everything from wildfire detection and material design to carbon monitoring and weather prediction. AI is helping us to respond to natural disasters, increase carbon storage and reduce wastage while optimizing production.
With the right direction, AI could form a critical backbone for the "unstoppable" clean energy revolution. For example, at last year’s annual UN climate summit, 198 countries signed the historic UAE Consensus, which promised to transition away from fossil fuels and triple renewable energy capacity by 2030.
It was one of the most ambitious climate pledges to date, one which will require technological advancements to fill current stopgaps and insufficiencies — balancing profit with progress.
To ensure AI is used to herald a new era of renewables and not entrench the power of Big Oil, we not only need Big Tech companies onboard, we need commitments and incentives to use AI to advance clean energy systems.
As Dr Sultan al-Jaber, the President of COP28, recently said, AI could be “the bridge [to] reduce emissions while also driving social and economic value everywhere.” He urged businesses to “collaborate to drive down the emissions of conventional energy [still needed] and on ways that AI can drive energy efficiency.”
For example, we can predict peaks and troughs in electricity demand and renewable generation with responsible, energy-focused AI. We can balance out electricity grids and automate systems such as lighting, heating, or cooling to operate more efficiently.
AI could also prevent waste — alerting us to faults within our energy networks, enabling us to intervene quickly and prevent further damage.
But to ensure AI is used to herald a new era of renewables and not entrench the power of Big Oil, we not only need Big Tech companies onboard, we need commitments and incentives to use AI to advance clean energy systems.
Climate change, a 'crisis multiplier'
One such solution is a supranational AI body with enough teeth to challenge these global corporations, monitor their emissions, and force them to mitigate the environmental impacts of their AI operations.
And, wild as it may sound, NATO, the unique transatlantic alliance between Europe and Northern America, is in a prime position to take up the reins.
Having adopted its first AI Intelligence Strategy in 2021, NATO has kept itself abreast of the technology’s risks and opportunities as they emerge. For example, it has incorporated AI into its own operations, piloting its use in areas including climate modelling and the analysis of satellite imagery.
Earlier this year, it began its development of a world-first certification standard that will help industries and institutions to ensure that new AI projects are compliant with international law.
Besides possessing the clout and recognizing the pitfalls of AI, NATO’s Secretary General has also called climate change “a crisis multiplier” and promised to reach net zero by 2050.
With its knowledge of AI technology and tunnel-visioned focus on safety and security, NATO can take the lead in establishing and enforcing mandatory rules for Big Tech corporations while resisting the siren call of deep-pocketed lobbyists campaigning for deregulation.
Ultimately, we cannot trust that individual state leaders will not prioritise AI’s generous profits over human interests. Already, we have seen the early signs of a dystopian future, characterised by surveillance, manipulation and the consolidation of power in the hands of a few.
Of course, this greed is hardly unique to AI. But never before have we birthed a technology that treads such a thin line between saving our world and driving it to self-destruction.
Manuel Matos dos Santos is Board Member for the Portuguese Atlantic Council, Community Manager at Closer Consulting, and former Secretary General of the Youth Atlantic Treaty Association (YATA) International.
At Euronews, we believe all views matter. Contact us at view@euronews.com to send pitches or submissions and be part of the conversation.
