Contrails can be avoided, just leave the yoke to Google

In short: Google Research, in a pilot (🥁) project together with United Airlines and Breakthrough Energy, has developed an AI algorithm that could empower airlines to significantly reduce the global warming impact of aviation. It uses real-time satellite imagery and weather data to recommend slight adjustments to a plane’s altitude in order to avoid the formation of contrails from exhaustion - the white clouds trails you see in the sky.

Deep dive: Contrails account for 35% of the industry’s global warming impact (wow!). That’s because the sod from the plane’s exhaust can trigger the formation of ice crystal in the upper atmosphere, that especially that night time have a greenhouse effect on heat radiation, heat that would otherwise be leaving the atmosphere.

The team at Google Research started by training a machine learning model to distinguish contrails from naturally forming clouds by providing pre-labelled (by themselves!) satellite images with marked contrails in them. By understanding the combination of factor present at the time of formation (plane altitude, weather conditions, etc.), the new algorithm can project likely zones of contrail formation in real-time based on the satellite imagery and weather data of the day.

Context: This is far from Google’s first stab at using machine learning to fight rapid climate change. Based on ML-powered understanding of how people move around in the world, last year they started providing eco-friendly route recommendations in Google Maps. And since 2020, they have been working with Electricity Maps to route compute-intense tasks in the Cloud solution to geopgraphic locations in the world with the lowest carbon intensity in real time.

Ever since declaring a “code red” in December 2022, Google has been firing on all AI cylinders, releasing scientific research, features for existing products, integrations, experiments and new stand-alone products, all powered by AI. Every move is part of its intense sprint to get back to being recognised and perceived as a powerful player in the AI domain.

Using AI for real-life, practical and more sustainable use cases might be one area where Google can more quickly gain an advantage over its competitors in Microsoft, OpenAI, Meta, Amazon and maybe even Apple. At least, it has the credibility in the field from products like Translate, Maps, Home and Meet.

My thoughts: It is vital that we repurpose and dedicate major technological advances to combat rapid climate change. Any attempt at that should be applauded. Still, I am always wondering about the balance between emissions connected with AI model training and the theoretical reduction that it would enable, if applied at scale. And it is that last caveat about adoption that I will be monitoring closely. If that happens fast and broadly, then this might be an example to inspire more great, regenerative applications of AI.

Look out for: I’d be super curious to see this line of thinking applied to other sectors and even other areas of transportation, e.g. cars.

One of the main selling points for a majority autonomous vehicle car fleet is the theoretical ability to avoid congestion and in the process waste energy. That requires real-time traffic data, which Google already has via Google Maps. Look for the Waymo’s, Uber’s and Cruise’s of tomorrow to see a mutual benefit in using a centralised 3rd party real-time data layer like Google (or Apple) to make their rides more efficient and less wasteful, simply by knowing where more cars in the network are and will be going.

Disclosure: I had to look up what the steering stick (fork?) of a plane was called. Yoke(s) on me.

🧠 Worthwhile

• Travel mindfully. A though-provoking essay in the New Yorker scrutinising the motivations of travel and tourism. I am definitely more pro-travel as a way to empathise with and understand cultures different from your local… but that requires that you seriously engage with it at the destination and uphold that intentionality with the journey. I wonder what kind of technology could support that goal.

• Point of no return for Google. Ryan Broderick takes the reader back to the early days of Google: PageRank, AdSense and Google Reader. In the end, a pretty clear picture is painted of a company that because of the vast competition pressure from both social media and LLMs might be repeating the same behaviour that undermined its own predecessor.

• Spirituality = 4th Bottom Line? Tom Greenwood, founder of Wholegrain Digital and a pioneer of the digital sustainability field, is regularly sharing observations about the difficult path to adoption of truly sustainable behaviour for individuals and businesses alike. I really appreciated this essay that argues that in order to meet current global challenges, we need to strengthen our inner foundation and improve our collective wisdom. YES.

🌤️ The bright side

This pleasantly plumb bird called a takahē was thought extinct, but is now returning to New Zealands wild, the Guardian writes. Now we just need to discover an isolated colony of dodo as well.

Recursion is a regular publication of essays about the relationship between technology and the Earth’s biosphere. Once in a while, it might contain sci-fi short stories. Written by Morten Bjørn Hallkvist.

If you enjoyed reading this, feel free to subscribe below.