Q&A: Bezos Earth Fund CEO on how AI could help climate and nature

The scale of the challenges facing the world in tackling climate change and protecting nature are not modest, from transforming how we power our lives to reversing deforestation.

Jeff Bezos, the Amazon founder and world's second-richest man, used a small slice of his fortune to create the $10 billion Bezos Earth Fund in 2020 to find and scale solutions.

In April this year, the fund announced a $100 million call for proposals for projects which use artificial intelligence tools to tackle climate and nature challenges.

Context sat down with the fund’s president and CEO, Andrew Steer, at COP29 in Baku to find out more.

This year you've launched the AI for climate and nature challenge. What's the thought behind this challenge, and how's it progressing?

These are early days for AI and its application, and we aren't sure where it's going to yield the most impact. Our interest is in finding really knotty problems that smart people are trying to solve but they simply don't have the capacity.

So far we've had 1,500 applications, and some of them look really interesting. The frontline environmental person, or researcher on food, or the person that's designing grids in Vietnam - it's the intersection of those people with the AI community that we are trying to catalyse.

What do you see as some of those potentially transformative uses of AI tools for climate mitigation and adaptation?

One is alternative proteins. If we keep eating proteins in the form of meat, and if the consumption of meat continues to grow at the rate projected, we will have a massive loss of nature. We will also fail to address climate change.

The problem is there are 200 million proteins in the world - in fact there's an infinite number. AI can play an amazing role in sorting through at a scale and a pace that would have been totally impossible otherwise.

Another one is creating smart grids in developing, emerging countries. AI can help you not only get all of your timing and sequencing right, but it also can tell you exactly what is the optimal way to expand your grid.

A third area for us is biodiversity conservation. AI can help hugely identify species, but it also can help recognise species that we do know exist, through sight, sound and even DNA.

For example, you think about invasive species. In Florida, the Burmese python was released a few decades ago, and now there are tens of thousands of them wreaking havoc. So they put all kinds of traps down, but most of those traps actually catch other animals, they don't catch the pythons. AI can, through sight recognition and even by scratching a little bit of the scales, trigger a trap just for that.

The flipside is the impact that the proliferation of data is having on the environment, especially from major firms like Amazon. Are you concerned about AI's footprint?

Yes, I think that's a big issue. There are two schools of thought: One is technology itself will help solve that, and even AI can help create a dramatic reduction in the need for such vast amounts of electricity, and it can even help create the renewable energy itself at a much cheaper cost.

So there's that optimistic side, but then there's another view that says this is going to overwhelm the demand for electricity and it's way too dangerous. The answer is we can't control the growth of AI obviously, we just need to make sure that we monitor that it doesn't do more harm to the environment than it does good.

We're exploring how valuable AI can be in getting us to where we need to get to by 2050, which will require exponential change. We're nowhere near on track, whether it's on nature or climate change.

The fund has been criticised for the amount of money put into carbon markets. Do you think AI and tech could create more transparency and more verification in those markets?

We don't put any money into carbon markets, we don't invest in carbon markets. What we have done is we have supported some of the standard-setting organisations.

We do believe that it is possible to have high quality carbon markets, and AI absolutely can help. I mean, we are financing satellites right now that can measure methane, and you could envisage in time that could be a legitimate player in carbon markets.

We also are investing a lot in the measurement of deforestation and reforestation. Now thanks to AI and the ability to manipulate large data and machine learning, we actually can figure out how much carbon is embedded in the trees that are falling and the trees that are growing.

It's still learning so it's improving the quality, but for the so-called 'carbon flux' one can figure out what's going on for every hectare on the planet from nature. That, in turn, would enable a lot of the abuses that have happened in carbon markets to be stopped.

Just like any emerging market, there have been abuses. You think about the development of a stock market, it took a long time. We can't afford to take that long, so AI can potentially help us get there much quicker.

This interview has been edited and condensed for clarity.

(Reporting by Jack Graham; Editing by Ayla Jean Yackley.)