Climate change is not just the most complex challenge humanity faces today, it’s also the most widespread. WHO data recently revealed that 3.6 billion people, almost half the world's population, are currently categorised as living in areas highly susceptible to climate change.

With direct costs of climate change set to rise to $2-4 billion per year by the end of this decade, where its causational annual death toll will likely hit 250,000, effective and rapidly deployable solutions are needed now. Artificial Intelligence, already transformative in industries as diverse as retail, energy production and mass-transit systems, clearly has a role to play in fixing our broken climate.

But how exactly is AI performing today in terms of boosting our collective response to climate change? And what is the price tag attached to this assistance?

Prediction: AI’s predictive power outstrips that of even the most informed humans, when the datasets are large and accurate enough. AI is now routinely employed in solutions designed to predict when and where climate disasters are likely to occur. In June this year The United Nations World Food Programme (WFP), Oxford University Physics Department, IGAD Climate Prediction and Applications Centre (ICPAC) and other groups joined forces in East Africa to create a new standard of weather forecasting empowered by AI. The current solution already offers highly accurate early warnings regarding rainfall and impending drought; a game-changer in disaster risk management, allowing relevant institutions to shift from reactive to preventative measures.

Detection: Assessing the scale of a problem is a prerequisite to solving it, as this ensures the right resources are assigned to the solution. AI is now instrumental in mapping large-scale climate phenomena such as deforestation and the melting of icebergs. An Edinburgh-based company Space Intelligence claims its AI solution has mapped more than 1.5 billion hectares of land across 30 countries from space using satellite data, to accurately gauge the severity of deforestation and the efficacy of subsequent reforestation efforts.

Collection: AI is not just a part of the planning process when mounting a climate response – it can be integral to the physical actions involved. Waste identification and collection is a highly complicated business, particularly when it’s floating across the world’s rivers and oceans. Non-profit organisation Ocean Cleanup (headquartered in Rotterdam, Netherlands) currently engages in multiple AI-based projects to monitor and actively collect ocean plastic waste streams. Ambitious in its long-term efforts, the group has calculated that, based on its past successes and the current level of AI-empowered technological solutions, it is now possible to clean up the Great Pacific Garbage Patch for a total cost of $7.5 billion.

Redirection: Solving the climate problems of past and present is essential, but so is correcting our mistakes and developing future industries that are entirely sustainable. AI is a core pillar of industrial decarbonisation efforts worldwide, as it uses the operational data of major carbon emitting sectors to find efficiency boosting measures and more radical transformation methods. A recent report from McKinsey stated that AI-driven technologies can now reliably help industrial enterprises reduce their CO2 emissions by up to 10% and cut energy costs by 10-20%.

While integral to producing insights that allow for a radical reimagining of our lives, living spaces, jobs and resource management, AI is also a very young technology with a lot of rough edges needing to be smoothed out.

A report last year from the London School of Economics (LSE) broadly categorised the risks of AI adoption in climate efforts as being threefold – its carbon footprint, necessary materials, and its capacity to raise inequality.

Carbon footprint: AI comes with a huge energy bill. A single ChatGPT search is around 25 times more energy intensive than a regular search engine query. While AI is constantly driving new improvements in energy efficiency, its benefits may not always outweigh its own CO2 emissions.

Materials:  Cobalt, lithium and tantalum and other rare critical minerals are necessary in the manufacturing and maintenance of advanced AI-based computing systems and support infrastructure. Again, their environmental benefits may outweigh the environmental cost of extracting these materials in the long term, but initially this is a significant cost factor to bear when the planet is already in crisis.

Inequality: As a broad-ranging technology with innumerable applications, AI is being developed across the world. However, heavier concentrations of research can be found in richer and more developed countries. Unless research institutions and companies have an expressed desire for open collaboration and/or philanthropy, it stands to reason that the benefits of AI research may also be concentrated in the developed world, rather than in more at-risk areas globally.

While the costs-to-benefits ratio of AI is still being hotly debated and contested across academia and global industry, the technology’s potential to bring about rapid transformative change is undisputed. Equally, concerns about its drawbacks have so far been insufficient to dissuade further experimentation. The race to develop world-shaping AI solutions is on, and there is no shortage of new contenders joining in every year.

As with business, AI has such a wide range of climate applications, the pace of its evolution is undoubtedly going to accelerate as a breakthrough in one area leads to faster progress in others. Given the collaborative, collective will to tackle climate change and transition to a sustainable climate future, AI is more than likely to play a starring role.