Is using AI worse for the climate than taking a flight?
The claim: “Using AI is as bad for the climate as flying.”
No, for ordinary chatting. A 1,000 km domestic flight emits about 246 kg of CO₂e per passenger by UK government factors (via Our World in Data), which include the extra warming of emissions at altitude. A short question at OpenAI’s 0.34 Wh on the world-average grid is about 0.16 g of CO₂. So the flight equals about 1.5 million questions, or about 0.8 million counting the flight’s CO₂ alone. Even at the highest published figure on a coal-heavy grid, about 1.04 g a question, it’s over 100,000. AI video and all-day coding agents use far more per task.
The bigger pictureAll the world’s data centres, of which AI is one part, used about 485 TWh in 2025. That’s about as much electricity as France uses in a year, and around 1.5% of the world’s. The IEA expects it to roughly double by 2030.
Why the claim misleads
It sets one person’s chatting against one person’s flight. The comparison that matters is the total.
What nobody knows yet
- There’s no solid public figure for heavy uses like coding agents running all day.
Sources (9)
- Our World in Data (UK government factors) Estimated
“A domestic flight emits 246 grams per kilometer.”
Quote checked on the page, 18 Oct 2026 - Our World in Data Estimated
“a "multiplier" of 1.9 is applied to aviation emissions to account for this.”
Quote checked on the page, 18 Oct 2026 - OpenAI (Sam Altman; average query, method not published) Company says so
“the average query uses about 0.34 watt-hours, about what an oven would use in a little over one second”
Quote checked on the page, 2 Oct 2026 - Ember, Global Electricity Review 2025 Measured
“Emissions intensity dropped by 2.3% to 473 grams of CO2 per kilowatt hour (gCO2/kWh)”
Quote checked on the page, 2 Oct 2026 - MIT Technology Review (one large open model, Llama 3.1 405B) Estimated
“an estimated 6,706 joules total, for each response”
Quote checked on the page, 2 Oct 2026 - Ember, Global Electricity Review 2025 Measured
“China's carbon intensity of electricity generation was 560 gCO2/kWh, down 4.1% from 2023”
Quote checked on the page, 3 Oct 2026 - IEA, Key Questions on Energy and AI (2026) Estimated
“electricity consumption from data centres roughly doubling from 485 TWh in 2025 to 950 TWh in 2030”
Quote checked on the page, 3 Oct 2026 - Ember, yearly electricity data (CSV, CC BY 4.0) Data row
Data row: France, 2025, electricity demand 480.6 TWh. Demand includes grid losses, so it is close to, not the same as, consumption.
Row checked against the published file, 6 Oct 2026 - IEA, Energy and AI (2025) Estimated
“Data centres accounted for around 1.5% of the world’s electricity consumption in 2024, or 415 terawatt-hours (TWh).”
Quote checked on the page, 3 Oct 2026
Hum’s sums
- 6,706 J ÷ 3,600 ≈ 1.86 Wh
- 246 g × 1,000 km = 246 kg of CO₂e per passenger
- 0.34 Wh × 473 g/kWh ≈ 0.16 g of CO₂
- 246 kg ÷ 0.16 g ≈ 1.5 million questions
- 246 ÷ 1.9 ≈ 129 g of CO₂ alone per km, so about 129 kg for 1,000 km
- 129 kg ÷ 0.16 g ≈ 0.8 million questions
- 1.86 Wh × 560 g/kWh ≈ 1.04 g of CO₂
- 129 kg ÷ 1.04 g ≈ 124,000 questions (over 100,000)
Cite this
Our World in Data (UK government factors), as quoted in Hum, “Is using AI worse for the climate than taking a flight?”, https://humearth.org/q/flight (quote checked 18 Oct 2026).
Cite the original source first; Hum is where you found it.
What do Wh, kWh and TWh mean?
Wh is a watt-hour: a 1-watt light left on for an hour. 1 kWh = 1,000 Wh, and 1 TWh = 1 billion kWh. Hum compares each number with something everyday so you don’t need the units.
Send this answer
One 1,000 km domestic flight is about 0.8 to 1.5 million short AI questions in CO₂ (UK factors via Our World in Data; OpenAI’s 0.34 Wh; world-average grid).