An Italian study reveals that generative AI is primarily used by younger men for information seeking, replacing traditional search engines.
Widespread access to powerful AI tools without instructions poses societal risks related to ethics, environment, and how we use technology.
A scientific paper analyzes how different prompt characteristics impact energy consumption during LLM inference using open-source models.
Research shows that the length of an AI's response, not the prompt, is the main driver of energy consumption, with variations between tasks and models.
Improve prompt sustainability by providing specific length and format constraints to guide the model and prevent verbose, energy-intensive responses.
The AI Watch browser extension helps you become aware of your prompt's environmental impact by estimating energy consumption and emissions for ChatGPT and Claude.
Adopt sustainable prompting by providing context, setting output constraints, and avoiding verbose language to help build an intelligent and sustainable AI-powered future.
Jobs that call for the skills explored in this talk.