TED Talks Daily We’re doing AI all wrong. Here’s how to get it right | Sasha Luccioni
99 snips
Oct 30, 2025 
Sasha Luccioni, an AI sustainability expert, reveals the staggering environmental toll of large AI models and massive data centers. She advocates for smaller, more efficient AI that minimizes energy usage while maintaining performance. Luccioni emphasizes the importance of curated training data to reduce misinformation risks and highlights practical benefits like enhanced privacy and accessibility. With innovative applications in climate monitoring, she calls for transparency in AI's energy use and collective action to promote sustainable practices in the industry.
AI Snips
Chapters
Transcript
Episode notes
Concentration Creates Environmental Harm
- A handful of large corporations are building massive data centers to chase superintelligence at the expense of people and the planet.
- Sasha Luccioni argues this concentration mirrors 'big oil' and creates unsustainable environmental and social harms.
Shift Power To Small, Local Models
- Reclaim power from big AI firms by supporting smaller, efficient models and local control.
- Prioritize models that use less compute and return governance to developers, regulators, and users.
Big Models Are Energy Inefficient
- Large generalist LLMs consume far more energy per simple task than small task-specific models.
- Luccioni's study found LLMs can use up to 30 times more energy answering straightforward questions.