The world’s average temperature is shifting at an alarming rate due to the rising emissions of greenhouse gases. At present, our planet is roughly 1.2°C warmer than it was before the industrial era. Unfortunately, due to the enduring impact of gases already discharged into the atmosphere and the challenges of abruptly stopping emissions, numerous experts believe that averting a temperature rise beyond 1.5°C is becoming an increasingly unattainable objective. But there is a ray of hope in our fight against climate change, thanks to the rapid advancements in artificial intelligence (AI), green information and communication technologies (ICTs), and robotics.
The International Telecommunication Union (ITU) is teaming up with partners worldwide to tackle this urgent issue. Initiatives like “AI for Good” are actively working to identify and enhance solutions where AI can assist us in identifying, adjusting to, and responding to climate change in various areas. This includes, but isn’t limited to, weather forecasting, energy preservation, and cutting emissions in sectors such as transportation, agriculture, and industry.
Will AI be helpful in addressing these pressing issues?
AI’s ability to contribute to the fight against climate change has been celebrated, and considering the current situation, the swift progress in AI is well-timed. This summer, mounting evidence suggests that the Earth is moving beyond mere warming to dangerously high temperatures.
Even though AI has generated a lot of buzz lately, it comes with a laundry list of concerns. These encompass its ability to spread misinformation, along with worries about bias, privacy, and security.
Moreover, researchers at the University of Cambridge in the UK have found that prejudice in the data used to train AI models might obstruct their ability to act as an impartial tool in addressing global warming and its impact on both the environment and human welfare.
Also read: Unleashing the power of AI Engine Optimization: Is it overtaking traditional SEO?
Best way to leverage AI for climate action strategies
Similar to many global biases, the issue can be traced back to a divide between the Global North and South. As most data is gathered by those with superior access to technology, the perspective on climate change’s effects becomes narrow. Consequently, biased AI has the potential to warp climate information, resulting in the most vulnerable populations suffering the most severe consequences.
In a paper titled “Harnessing human and machine intelligence for planetary-level climate action” published in the highly regarded journal Nature, the authors acknowledge that “using AI to account for the continually changing factors of climate change allows us to generate better-informed predictions about environmental changes, allowing us to deploy mitigation strategies earlier.”
They argue that this is still one of the most hopeful ways AI can contribute to climate action strategies – but only if the datasets used to train these systems are truly representative on a global scale.
Dr. Ramit Debnath, the lead author, and a Cambridge Zero Fellow, emphasized that if climate change data predominantly reflects the perspectives of highly educated individuals from top-ranking institutions in the Global North, artificial intelligence will inevitably view climate change and its solutions from their limited viewpoint.
“When the information on climate change is over-represented by the work of well-educated individuals at high-ranking institutions within the Global North, AI will only see climate change and climate solutions through their eyes,” he said.
Biased data can be a challenge for AI
Conversely, those who have restricted access to technology and reporting resources will experience exclusion from the digital repositories that AI developers rely upon.
Professor Emily Shuckburgh, a co-author of the paper, highlighted the inherent bias and contamination present in data, presenting a formidable obstacle for AI due to its heavy reliance on digital information. She stressed the crucial importance of recognizing this disparity in data and using it as a starting point to tackle the issue, ultimately leading to the creation of more dependable AI-driven climate solutions.
“No data is clean or without prejudice, and this is particularly problematic for AI which relies entirely on digital information. Only with an active awareness of this data injustice can we begin to tackle it, and consequently, to build better and more trustworthy AI-led climate solutions,” Shuckburgh said.
Also read: Can AI help you break free from virtual meeting overload?
AI models with human inputs could work
The authors are in favor of AI designs that include human participation, as this can contribute to establishing a worldwide knowledge network to bolster climate action. These design approaches can actively assist in the endeavors to lessen the impact of climate change and adapt to it, all while reducing the data biases often linked with AI training datasets.
The research highlights the importance of recognizing digital inequalities and injustices, especially in the realm of machine intelligence, particularly when AI is used to address global health concerns like climate change.
As per the authors, failing to address these issues could have catastrophic results, including societal collapse and endangering the stability of our planet. This could potentially throw a wrench into all attempts to mitigate climate change.