Climate Action in a world of AI

While some fear-based messaging can be motivating in the short-term, research shows that it backfires when trying to inspire long-term action.


Introduction

Good morning, I'm Lee Mallon, the Chief Technology Officer (CTO) of Humanity. Today, I'm revisiting my presentation from COP28, focused on climate action in a world of Artificial Intelligence (AI). The original talk, constrained by time, is now extended to delve deeper into this crucial subject.

Humanity's Mission

Humanity, an organisation I represent, is dedicated to advancing storytelling within the broad scope of driving human progress. With over two decades of experience in technology, particularly in AI, data, and natural language understanding, my focus has been on data-driven storytelling. This approach is crucial for creating personalised narratives that resonate with a global audience, making technology and impactful causes more engaging.

My passion has led to the creation of a book titled "DKR: Dynamic Knowledge Rendering," set for release by Christmas. It's a testament to our commitment to making complex information accessible and engaging. For those interested, I'm offering free copies upon request.

Inspiration not Fear

If you’re like me, most of what you hear about climate change in the media is negative and dire. News reports barrage us with statistics about rising emissions, extreme weather events, collapsing ecosystems, and so on. There’s a lot of doom and gloom around missing our chance to limit warming to 1.5°C. While some fear-based messaging can be motivating in the short-term, research shows that it backfires when trying to inspire long-term action.

Specifically, human behaviour expert named Lee Groombridge has found that while fear initially jolts people into caring, it cannot sustain decade-to-decade action the way positive visions of the future can. Climate communications need a major reset. We need to showcase the benefits and possibilities of a net zero emissions world. What would our cities look like powered by 100% clean energy? How would air quality improve? What jobs would be created? Painting a compelling vision of a thriving, sustainable future gives people and communities something to strive towards collectively.

AI as a Tool for Climate Communication

AI, particularly large language models like ChatGPT, trained on vast internet data, have revolutionised our interaction with technology. These models understand human language, allowing us to converse naturally with machines. This advancement has significant implications for climate communication.

In recent years, AI systems have made extraordinary progress in language processing thanks to a class of algorithms called Large Language Models (LLMs). LLMs like ChatGPT and Claude are “trained” on huge datasets of online text and dialogue to develop common sense reasoning and communicate fluently through text or voice interfaces. They have moved AI from narrow pattern recognition tasks like playing chess to impressively human-like conversation and content generation.

I demonstrated what LLMs can do for climate data by uploading the 100-page UNEP report into ChatGPT. Within minutes, I could ask ChatGPT to summarise key findings, suggest potential questions, create charts and social media graphics, generate quiz questions for students, and more. Anything I requested, ChatGPT could produce on the fly by extracting and repurposing relevant passages of text and numbers from the report.

Now anyone can access a powerful AI assistant to make sense of dense climate resources. Asking your own questions and getting customised explanations tailored to your background knowledge is often more useful than pre-written summaries from experts. And visual communication through charts, images, videos, and even poems and music can convey insights more memorably and emotionally than blocks of text. Long story short: there has never been an easier way to get climate-literate.

Hyper Personalisation: The Future of Storytelling

Part of why climate messaging often falls flat is that critical data and insights are trapped in long, dense reports that few people have time to parse through. For instance, in the lead up to COP28, numerous organisations published flagship reports over 100 pages in length, packed with charts, analysis, and projections. Even for climate experts, holding all the key takeaways in one’s head is a challenge.

But what if there were a way to rapidly simplify, summarise, and visualise the treasure troves of data in these reports for wider audiences? What if the insights could be accessed through natural conversation instead of demanding hours of focused reading? AI technology has advanced rapidly enough to turn these possibilities into realities.

While AI democratisation of climate data is game changing on its own, truly next generation communication uses personalised narratives powered by machine learning. I gave a sneak preview by using ChatGPT to generate a customised climate vision video for my volunteer participant, Gerry. Based on a photo of Gerry, a voice clip, and a couple sentences about his passion for a sustainable Kenya, ChatGPT wrote an inspirational script about Kenya’s renewable energy transformation toward 2047. Complete with AI-generated music and visuals, this positive vision video resonated much more than reciting statistics from the latest IPCC report from a future, older Gerry reflecting back at him the change he has made.

Now imagine millions of Gerry's all with hyper-personalised content. AI can scale the creation of personalised climate stories focused on what different communities care about most. For Miami, it may focus on adaptation that were made in the face of sea level rise. For Stockholm, it may highlight the health co-benefits of going fossil fuel free. And for Mumbai, it may envision how green jobs have lift millions out of poverty. Moving beyond one-size-fits all messaging is key. Research on psychological biases shows that when information resonates with what people value, they become much more supportive of climate action.

Our collective future depends on getting this right. If misinformation obscures the truth and data-dense reports collect virtual dust, fear and denial win. But if AI can turn climate insights into inspiring visions of flourishing communities. The technology is here, we just need the ambition and creativity to use it wisely. I’m hopeful that the seeds planted at COP28 will grow into a new era of climate storytelling - one that brings out the best of human compassion through the power of AI.