OpenAI wants to give us 5% of its success. It's a bad bargain
Ben Werdmuller argues that OpenAI's offer of a 5% stake in its success in exchange for public data is a bad bargain, as it undervalues the collective contribution of humanity's knowledge. He critiques the framing as a trick that would legitimize the exploitation of public data for private profit without meaningful compensation or governance.
Background
- OpenAI is currently restructuring from a capped-profit "capped" company into a traditional for-profit business, a move that could unlock enormous value for investors and executives.
- As part of this transition, OpenAI has proposed giving the public (via a "Public Benefit Corporation" structure or similar) a 5% stake or interest in its future success — framed as a philanthropic gesture.
- The article argues this is a bad deal: 5% is far too small a share given that the company's AI models were trained on vast amounts of publicly created data (text, images, code from the open web), and its ongoing success depends on continued public contribution and trust.
- The core tension: private companies are capturing most of the value from AI systems built on collectively created knowledge, while offering the public only a token stake in return.
- This is part of a broader debate about "data as labor," AI companies' relationship to the commons, and whether AI benefits will be broadly shared or concentrated among a few.