In short
Reduce first, and treat offsets as optional extras rather than a way to make AI use "carbon neutral". For most people the footprint of everyday text AI is small, and cutting the heavy uses, such as image and video generation and endless regeneration, does more than buying credits. Standards for software carbon do not let offsets reduce the measured footprint.
Steps
- Cut the heavy uses first, such as image and video generation and repeated regeneration.
- Estimate roughly how much AI you use before deciding anything else.
- If you buy offsets, treat them as a contribution, not as cancelling your use.
- Be wary of apps or services that call AI use "carbon neutral" because of offsets.
Reducing beats offsetting. An offset pays for a reduction somewhere else; it does not change the electricity your AI use draws or the hardware it needs. For most people, the better move is to cut the uses that cost the most.
Is my AI footprint big enough to offset?
For occasional text questions, it is small. On OpenAI's figure, an average ChatGPT query uses about 0.34 Wh, about what a high-efficiency light bulb uses in a couple of minutes. Heavy image and video generation, or long documents pasted into chats, are much larger. To get a rough number, see work out your own AI carbon footprint.
Why are offsets a weak answer?
- They don't change the measured footprint. The Green Software Foundation's Software Carbon Intensity specification does not let offsets, renewable certificates or similar market-based measures reduce a score.
- Quality is hard to check. An individual buyer rarely has a way to confirm what a project delivers.
- Claims are tightening. In the EU, Directive (EU) 2024/825 restricts environmental claims to consumers that cannot be proved, including claims of neutrality based on offsetting.
What makes the biggest difference?
Generating fewer images and videos, refining rather than regenerating, and keeping devices longer. If you then want to support climate projects, do so as a contribution, without describing your AI use as "carbon neutral".
To judge claims made by companies, see honest green claims and spot greenwashing in AI marketing.
Sources
- Software Carbon Intensity specification, Green Software Foundation (2024)
- The gentle singularity, Sam Altman (OpenAI) (2025)
- Directive (EU) 2024/825 on empowering consumers for the green transition, European Union (2024)
Related
How do I work out my own AI carbon footprint?
A rough personal AI footprint: count your text prompts, images and long tasks, multiply by published per-request figures, and treat it as a range.
How do I put my AI energy use in perspective?
How one AI question compares with everyday energy use, why images, video and long documents change the picture, and why the global total still matters.
How do I spot greenwashing in AI marketing?
Warning signs of greenwashing in AI products and services, from vague 'sustainable AI' labels to offset-based neutrality claims, and the questions to ask.
Honest green claims about AI
How to avoid greenwashing about AI: what you can and can't claim, rebound effects, honest reporting, and the UK, EU and US rules on environmental claims.