Using computing to fix climate change
If you'd like to use computer science to help mitigate climate change, then check out these resources:
- ClimateChange.AI - All about using AI to mitigate and adapt to climate change. Check out their excellent 2019 paper and sign up to their newsletter! Also check out Paul Strobel's six-part illustrated blog series about the paper.
- Open Climate Fix - OCF is the non-profit that I co-founded. Entirely focused on using open science (especially machine learning) to mitigate climate change. Please sign up to our newsletter and consider adding your name to our list of volunteers!
- Matter More - 'Use your superpowers in AI and data science to reverse climate change - Discover exciting new initiatives and work opportunities.'
- Sliced - 'We help you find opportunities that have a direct impact on reversing climate change.'
- ClimateAction.Tech - 'A global community of tech professionals using our skills, expertise and platforms to support solutions to the climate crisis.'
- Bret Victor's blog: What can a technologist do about climate change - 2015
- Dynamically Typed newsletter - bi-weekly newsletter about AI-powered products and ML research that also features new projects and resources for using AI to tackle the climate crisis in every edition.
If you know of other resources, then please add a comment below or, even better, submit a pull request to improve this list!
The challenge of reducing emissions at scale
One quick rant: All the climate cares about is the scale of emissions reductions. For the climate to 'notice', we need to reduce emissions by at least a million tonnes of CO2e per year, and ideally billions of tonnes per year.
To achieve scale we need to jump over multiple hurdles:
- Have a great research idea.
- Talk to industry to check that your idea solves a real problem. (I don't know about you, but I find it all to easy to fixate on solving 'toy' problems. It's essential that we focus on solving 'real' problems.)
- Build a proof-of-concept that demonstrates emission reductions in a controlled environment.
- Build a product (or get your idea implemented into an existing product) that reduces emissions and solves problems that industry cares about, at a price they can afford, and in a way which fits easily into their existing systems.
- Persuade industry to adopt the product / idea. (Marketing, support, iterating on the product after receiving critical feedback, etc.)
- Become financially sustainable.
- Reduce emissions! (Even if you succeed at all the previous steps, that's no guarantee that you'll reduce emissions at scale. Maybe there's a rebound effect? Maybe there are some other unintended consequences?)
Projects tend to attract most funding and media attention after completing a proof-of-concept. But we need to get all the way to the final step before we can be confident that a given intervention will actually reduce emissions at scale. The systems that we're trying to change (such as the energy system) are enormously complex. Demonstrating success in a small research setting does not guarantee success at scale (unfortunately!).
Good luck! It's gonna be tough. But we can't fail. To paraphrase Greta: If we fail, future generations will never forgive us. (Sorry to be so gloomy!)
But it's also technically exciting work (and actually quite good fun!) and there's a great community of people using computer science to mitigate climate change! So, dive in - we need your help!