Archive — US — techcrunch
Meta is paying to peek at how you use their latest AI model
Meta is incentivizing users to share their data with its AI model, Muse Spark, by offering significant discounts—95% for users who contribute their prompts and outputs. This move comes as Meta seeks vital training data, having previously faced criticism for its data collection methods. The aim is to enhance AI capabilities while navigating the reluctance of companies to share sensitive information.
Key points
- Meta introduces Muse Spark AI model with a pricing model rewarding user data sharing.
- Users can save 95% on token costs by sharing prompts and outputs.
- Standard prices are $1.25 for input tokens and $4.25 for output tokens; contributor prices are 10 cents and 20 cents respectively.
- Meta's previous data collection initiative faced criticism and was paused.
- Historical context shows reluctance from companies to share proprietary data despite cost benefits.
- The contributor model is designed to encourage firms to evaluate their data's proprietary nature.
Why it matters
- The trend of companies like OpenAI and Anthropic lowering prices for AI services reflects a competitive landscape in AI.
- In 2025, the focus has shifted towards optimizing AI model training with user data and improving AI applications in various industries.
- Similar efforts in the AI sector reveal a broader industry trend towards incentivizing user participation for model training customization.
Meta's approach to encourage data sharing via discounts could reshape how companies interact with AI providers, balancing data privacy with the need for improved AI model performance as competition intensifies in the AI industry.