Google is using AI to avoid aircraft contrails. What it means for Malaysia

A white line behind an aircraft can look harmless, but some contrails persist, spread into thin cloud and trap heat. Google is now taking its AI forecasting work from airline-level tests into Operation Blue Skies, a UK-backed trial across the Shanwick Oceanic airspace in the eastern North Atlantic.[1][2]

The announcement matters to Malaysia as an example of AI moving into day-to-day flight operations, not because the system is already running here. Neither Google nor the Met Office named a Malaysian airline, the Civil Aviation Authority of Malaysia or the Ministry of Transport as a participant. That boundary is important: a promising foreign trial is evidence to study, not proof of local readiness.[1][2]

Contrails form when aircraft exhaust meets very cold, humid air at cruising altitude. Some disappear quickly. Others can remain for hours and spread, reflecting some sunlight but also trapping heat that would otherwise leave the atmosphere. The balance changes with location, time and weather, which is why forecasting is central to the trial.[2]

Google's role is to forecast contrail-sensitive regions using AI, while machine learning and satellite imagery help track and verify what happened. NATS will coordinate air traffic operations and safety work; the Met Office is developing forecasting capability; and researchers at Imperial College London and the University of Cambridge will independently evaluate outcomes. Controllers would use normal procedures to make targeted altitude changes for selected flights.[1][2]

Operation Blue Skies is planned for 30 months, with one operational window in winter 2026/27 and another in winter 2027/28. Google said around 10,000 flights from hundreds of cities would pass through Shanwick airspace during trial hours each year. Only a small percentage that would otherwise cross sensitive air is expected to be slightly diverted.[1][2]

There is earlier evidence, but it should not be mixed with the new trial. In 2023, Google reported that American Airlines pilots using its forecasts reduced contrail formation by 54% across 70 test flights, while the avoidance flights burned about 2% more fuel. Google estimated the fleet-wide fuel effect could be much lower if only the highest-impact flights changed. Those findings were an initial airline test, not an airspace-scale result and not a Malaysian result.[3]

Malaysia's Aviation Decarbonisation Blueprint, launched in September 2024, groups its pathway around aircraft technology, operational improvements, sustainable aviation fuel and market-based measures. Its short-term scenario assigns a 2% to 2.5% reduction range to operational improvement between 2024 and 2029. The blueprint does not make Operation Blue Skies a local programme, but it creates a policy home where validated operating measures could eventually be assessed.[4]

CAAM has also run a separate direct-route trial in the Kuala Lumpur Flight Information Region. That programme sought to shorten track miles and reduce carbon dioxide using existing waypoints. Contrail avoidance solves a different problem: it may deliberately change altitude, and sometimes add fuel burn, to avoid non-carbon-dioxide warming. A Malaysian evaluation would have to measure both effects instead of assuming that the shortest path is always the lowest-climate-impact path.[5][3]

Malaysia's tropical atmosphere, regional routes and airspace constraints differ from the winter North Atlantic setting. A serious local pilot would need forecast skill measured against satellite observations, safe integration with air traffic control, clear rules for choosing flights, and accounting that compares avoided contrail warming with additional fuel and carbon dioxide.[1][2][4]

It would also need transparent reporting. Useful evidence would show how many flights received a recommendation, how many followed it, whether predicted contrails actually disappeared, the added fuel and delay, and the net climate result. Until the UK trials report those outcomes, the strongest Malaysian conclusion is modest: this is an operational experiment worth watching, not a ready-made product to announce as deployed.[1][2]