A recent large-scale trial demonstrates that artificial intelligence can effectively reduce the climate impact of air travel by rerouting flights to avoid creating persistent contrails. The study, involving over 2,400 American Airlines flights between the US and Europe from January to May 2025, showed an overall reduction of 11.6% in contrail formation when pilots were provided with AI-optimized flight paths. For the flights that did take the AI-suggested routes, the reduction in visible contrails reached an impressive 62%.
The Problem with Contrails
Aircraft contrails – the white streaks left behind by planes – aren’t just harmless water vapor. They trap heat in the atmosphere, and recent research suggests they may contribute to warming more than the carbon dioxide emissions from jet engines. This is because the ice crystals in contrails form more readily in certain atmospheric conditions, and these conditions can be predicted using detailed weather models. Avoiding these “contrail-prone” regions is key to reducing aviation’s climate footprint.
How the Trial Worked
Researchers at Google, led by Dinesh Sanekommu, integrated their AI contrail-forecasting tool into American Airlines’ flight-planning software. Flights heading eastward (where contrails have a more pronounced warming effect at night) were randomly assigned to either a control group (no AI suggestion) or an experimental group (AI-optimized route available).
While dispatchers in the experimental group could choose the AI-suggested route, only 112 out of 1,232 flights actually did due to practical constraints like cost and safety. Nevertheless, even with partial adoption, the overall impact was significant.
Key Findings and Implications
The AI analysis of satellite imagery confirmed the reduction in contrail visibility. Flights following the optimized routes reduced their warming effect by 13.7% across the entire group, and a striking 69.3% for those that actually used the AI-suggested paths. Importantly, there was no statistically significant difference in fuel consumption, suggesting that contrail avoidance can be achieved without increasing operating costs.
“This validated the thesis that if we can safely integrate this into flight planning, it’s a scalable way to consider contrail avoidance,” says Sanekommu.
Future Outlook
Experts like Edward Gryspeerdt at Imperial College London call this “the best you can do” with current tools. Scaling up to a 60% reduction across all flights may be unrealistic given logistical complexities, but even a 10% reduction in contrails represents a substantial step towards more sustainable aviation.
The success of this trial demonstrates that AI-driven route optimization is a viable strategy for mitigating the climate impact of air travel. Further refinement of these tools and broader industry adoption will be crucial to realizing the full potential of contrail avoidance.
