Will AI save or break the planet?
This article originally appeared in I-M Magazine
The dominant rhetoric around artificial intelligence (AI) is that humans are forging the agents of their own destruction. Since Karel Čapek’s robotsfirst eliminated mankind in the 1920s, headlines have continually warned of technological dangers to come: from the real life terminatorto dangerous ‘digital superintelligence’, and mass job losses.
But there are growing indications AI is not the beginning of the end.
In addition to research that shows AI could create more jobsthan it displaces, there is evidence tech innovations could address a more fundamental and pressing issue: saving our planet.
We broke our world
In 2018, the UN Intergovernmental Panel on Climate Change (IPCC) released the chilling key finding of its global warming report: we have a dozen years to cut emissions or face large-scale implications. According to the IPCC, the planet is already 1C warmer than pre-industrial levels and likely to reach 1.5C between 2030 and 2052 if current rates continue. This matters because a further rise of 0.5C could significantly increase the risks of extreme weather such as drought and flooding, and poverty related to climate shifts. Loss of insect habitat, for example, may reduce crop yield through low pollination.
Unsurprisingly, the report has driven wider awareness of the looming potential crisis and damage to the ecosystem. Its call for “urgent action” has inspired both protestsand the creation of the ‘Katowice rulebook’, which details how the 2015 Paris climate accord should be enacted to stop temperatures exceeding the 1.5C bar. And with fresh studies showing global warming has heated oceans to the equivalent of one atomic bombper second over the last 150 years, interest in world-fixing solutions is higher than ever.
So, could AI be one of them?
While it’s true the raw materials needed to build smart tools do have an impacton the environment, AI developments could deliver greater long-term benefits.
Tech for humanity: life-saving AI
Firstly, using AI to control negative environmental impact doesn’t mean we will soon be hearing echoes of those immortal words from A Space Odyssey as machines take the wheel. Rather, it will allow us to achieve a comprehensive understanding of global problems, as well as the means to meet challenges quickly and efficiently.
Let’s take a look at the leading emerging applications:
- Unlocking new resources
Harm done by materials such as plastics is a hot button issue, with estimates citing the total volume of waste plastic at 6.9 billion tonnes. That’s not to mention the extraction of resources such as nickel and cobalt to make lithium-ion batteries for smartphones. But AI is also helping to offset the carbon footprint of tech through new materials.
For instance, researchers using the Open Quantum Materials Database (OQMD) have discovered metal oxides that could enhance solar cell efficiency: boosting its capacity to harness energy from sunlight and decreasing dependence on non-sustainable sources. Meanwhile, AI has enabled scientists collaborating with Northwestern University to discover three new models for constructing metallic glass with the help of automated algorithms, meaning they are a step closer to producing viable alternatives to steel.
Plus, tech giants are also leveraging AI large-scale data capabilities to restrict energy consumed by data analysis itself. Perhaps the best example of this is Google’s DeepMind, which drew huge stores of past data — covering factors such as temperature and pump speeds — to improve streamline cooling systems at all organisational data centres in 2016, and successfully reduced the energy required to operate them by 40%.
- Putting a lid on pollution
Providing much-needed light at the end of the climate change tunnel, researchfrom the World Economic Forum (WEF) and PwC has identified ways tech can be a force for good as part of the ‘AI for earth project’. Primary among them are uses to minimise pollution.
To start, it’s predicted AI autonomous vehicles will fuel a move towards mobility-on-demand, meaning lower levels of car ownership, usage, and emissions. Following on from this, intelligent systems are also expected to assist in running smart cities, where real-time data facilitates coordination of driverless fleets — via eco-driving and traffic optimisation algorithms that curtail greenhouse gases — and energy allocation. For example, self-learning tech could tap data about energy distribution to establish when and where power is typically needed, such as TV ‘pickup’ times when audiences collectively boil the kettle during ad breaks. It will then ensure renewable energy can meet demand by organising efficient storage, loading, pricing, and trading. In other words, it will improve the reliability of clean energy and allow it to compete with fossil fuels.
Then there is the concept of an international AI-based environmental console. Created using an open application programme interface (API) and constantly fed on data from all corners of the world, such a control panel could provide one central point from which multiple users can monitor and manage shared climate goals. This might include insight about rates of air pollution, crop production, or update natural disaster responses.
- Protecting more than ourselves
A final yet equally vital implementation for the master console is securing the survival of other species. Protecting various habits is not only important to maintain resources humans depend on — see what beesdo for us — but also safeguard the delicate balance of the ecosystem. As a result, it makes sense that the WEF foresees AI systems also tackling illegal deforestation, water extraction, fishing and poaching, in addition to studying and saving wildlife with guidance from machine learning models.
What will the future hold?
While the path ahead is unclear, AI has the potential to make it smoother and more sustainable: collating data that gives people and authorises the ability to appreciate the consequences of each action, and make better decisions. So, to an extent, AI could save the planet — but it can only do so effectively in combination with our experience and desire for change.
Without a foundation of human insight, AI has nothing to go on. This means the fate of the globe rests on our ability to blend findings from our scientific endeavours, intelligent tools, and a desire to preserve it for every future generation, not just the next.