Can AI help us tackle the climate crisis? We ran a hackathon to find out.
(Answer by a curious human, not a smart-arse bot)
Why ask the Q?
You don’t get many more buzzwords in a sentence than: AI, climate and hackathon. But we weren’t putting this event on to ride the ‘coming wave’. We genuinely wanted to understand the hype (and fear) around AI and whether the entrepreneurially minded amongst our connections were able to quickly put this new tech to use in the climate era. Why? Because despite the growth of incredible impact businesses and mass activism, the evidential breakdown of earth’s boundaries and climate injustice urgently requires more rapid, scalable solutions than middle class conscious consumption and flip-flopping Government policy.
How did the hackathon come about?
A linkedin post from UK tech-queen Sarah Drinkwater & community builder, Victoria Stoyanova, summarised their successful AI hackthon and helpfully shared a ‘playbook’ that led to a Q from a few us: why not try this format for tackling climate challenges?
Katie McPhee responded. We had a call. We decided to try and put on a climate focused AI hackathon. We called it EarthAI and put out the word.
What happened at the hackthon?
We kicked off on Friday afternoon asking each other ‘why are we here?’ It’s always good to understand each other’s motivations.
“I’m here because I’ve seen the devastating impacts of climate on farmers in Africa and I want to help.”
“I’m excited about how AI can really make a quick, big difference.”
Then a couple of mini-talks from people who work in this space: Dr. Caroline Chibelushi from Innovate UK KTN showed us how AI ideas get traction, followed by Stephen Brittain from Insurtech Gateway who shared how venture building in the insurance industry is embracing this technology to tackle climate.
20+ ideas were shared (a lot) which we then organically shuffled down to a dozen around which teams formed. This was a surprisingly fun and easy process. A lesson for engineering serendipity in decision making and recruitment?
Then came the action: Teams collected their OpenAI credits and started mapping out their idea. Fuelled on Pizza Pilgrims (this is a hackathon after all), Dash Drinks, Tribe Bars & Wednesday’s Domaine…the teams spent Friday night and the whole of Saturday building. Supported by mentors in-person and online, problem-solutions were refined and pulled together into a coherent version 1. There was no ‘stay up all night’ coding for this hackathon… we need balance people (or we’re getting old).
At 4pm we gathered for ‘Pitch Time’ — each team (who had met less than 24 hours earlier) had 8 minutes to present and demo their idea and take Qs from the judges.
Three winners were awarded a collection of prizes to give their projects more momentum beyond the weekend.
These are their ideas.
5can
Problem: Lack of transparency in the voluntary carbon market.
Solution: Leverage LLMs to automate data structuring. Analytical insights highlight projects with the highest emission reduction.
ELPS Energy
Problem: Need to improve carbonization of housing through retrofitting.
Solution: Develop an AI platform guiding homeowners from assessment to aftercare, promoting a superior supplier marketplace. Utilize public energy certificates to estimate energy consumption, cross-referencing various sources and employing feature recognition.
Switch EV
Problem: Challenges in making personalised recommendations for electric vehicles.
Solution: An EV chatbot offering tailored advice on switching to electric cars, considering user specifications such as location and range needs. Monetisation is achieved through partnerships with EV providers.
Cli-move — Winner: People’s Choice
Problem: Uncertainty surrounding climate migration and associated economic risks.
Solution: Agent-based modelling determines logic from input data, merging population migration datasets with climate risk datasets to produce insights.
Crop Uncle — Winner: High Impact
Problem: Many small-scale farmers, responsible for feeding 70% of the world, lack adequate climate change support, especially underfunded female farmers.
Solution: Implement a voice recognition and translation hotline offering climate resilience advice. The model is trained using diverse datasets, including research and location-specific information.
ClimbAlign
Problem: Consumers often struggle to make sustainable purchasing decisions.
Solution: A browser tool provides sustainability scores for products and suggests eco-friendlier alternatives. The system utilises Gen Ai to amalgamate live data from multiple sources.
Irrigation AI — Winner: Best use of AI to tackle Climate
Problem: Irregular climate patterns makes it hard to ensure water is being used efficiently and crops are being adequately nourished.
Solution: An integrated AI model evaluates 10 meteorological, soil, and plant visibility parameters. Images of leaves can be uploaded for further diagnosis. The system identifies water presence, optimal irrigation levels, and crop health indicators to guide cultivation for farmers.
Cleansplain
Problem: Greenwashing is on the rise, and incoming regulations will penalise companies practicing it.
Solution: A tool detecting three levels of greenwashing, targeting companies with ESG divisions communicating their climate impact.
Natural Elder
Problem: Forest fires cause massive destruction and dried wood — yet 72% of forest walkers would be willing to gather dead vegetation to mitigate forest fire risks.
Solution: AI directs these individuals to high-risk zones, with gamification techniques tracking and rewarding preventative actions.
Green Career Coach
Problem: Guiding young individuals towards green vocational careers.
Solution: A chatbot assesses users based on education and interests, directing them to environmentally-focused professions, emphasizing the increasing role of blue-collar jobs in the green economy.
So, can AI fix the climate and nature crises?
I’m in no position to answer this question with any authority. I don’t think any human is. But…… from this weekend’s little experiment the following is clear:
1. Smart, creative, compassionate people are motivated to work transformative solutions to our biggest challenges.
2. With the right expertise and tools, AI solutions can be rapidly created, presenting compelling proofs of concept. This demonstrates AI’s potential to provide immediate answers and strategies, especially when time is of the essence in combating climate challenges.
3. The hackathon format is as good as any short event/workshop/bootcamp I’ve seen to learn fast, build new teams and create something that might just be the start of everything. I wish we had these at school.
Will you do this again?
We’re hosting a pitch night in November (limited tickets available) to showcase the first batch of ideas (and more) to an audience who could help them scale.
At Raaise we’re helping climate focused founders organise and manage their fundraise so they can get on with building the solutions we so desperately need to succeed.
If you would like to partner with us and scale this up, then we’re ready to chat.
Hackathon Thank You’s
Huckletree: for an awesome venue + team
Insurtech Gateway: for an inspiring kick-off and mentorship
OpenAI: for credits to build and award (and tools to change our world)
Raaise: for powering climate startups through their fundraising rounds
Upwards Studio: for mentoring and powering innovators
Terra.do: For climate career education
Climate Connection: For gathering London’s community of builders.
EarthAI Mission Control: Vessy Dragusheva & Joe
Mentors & Judges: Chris Ballard, Ilona Ludewig, Dhen Padilla, Stephen Brittain, Jana Tian, & Henry Blanchard.
Volunteers: Zoe, Belinda, Remy, Alex, Anca
Sim Visuals: for awesome shots
