These AI chatbot questions cause most carbon emissions, scientists find

illuminem summarises for you the essential news of the day. Read the full piece on The Independent or enjoy below:

🗞️ Driving the news: A new study from Hochschule München University of Applied Sciences reveals that complex queries requiring AI chatbots, like OpenAI's ChatGPT, to perform logical reasoning result in significantly higher carbon emissions
• The research finds that answers involving abstract reasoning, such as philosophy or abstract algebra, produce up to six times more carbon emissions than simpler queries like high school history
• The study emphasizes the environmental impact of AI models, particularly those requiring intensive reasoning processes

🔭 The context: The study compares 14 AI models' carbon emissions using 1,000 standardized questions
• It identifies a clear correlation between the complexity of a question and the amount of energy consumed by the AI to generate a response
• The findings underscore the growing concern over the sustainability of AI technologies, especially as their use becomes more widespread in various sectors

🌍 Why it matters for the planet: This research highlights an often-overlooked environmental cost of AI technologies
• As AI models evolve to provide more accurate and nuanced responses, their carbon footprint increases, raising questions about their long-term sustainability
• The study emphasizes the need for both users and developers to consider how AI can be optimized to minimize environmental impact, especially given the growing global focus on reducing carbon emissions

⏭️ What's next: The study calls for AI users to adjust the types of questions they ask to reduce energy consumption and carbon emissions
• It also signals a need for future AI model development to prioritize sustainability alongside performance
• Researchers hope that these findings will encourage more environmentally conscious usage patterns and spur innovations in AI efficiency

💬 One quote: “We found that reasoning-enabled models produced up to 50 times more carbon dioxide emissions than concise response models.” — Maximilian Dauner, study author

📈 One stat: Reasoning-enabled AI models produced up to six times more carbon emissions than simpler models in response to complex queries

See on illuminem's Data Hub™ the sustainability performance of OpenAI and its peers Anthropic and Deepseek

Click for more news covering the latest on carbon and green tech