AI Dependency Mirrors 1980s Software Freedom Crisis, Warns Open-Source Advocates

The free software movement, born in the 1980s after Richard Stallman encountered a locked-down Xerox printer at MIT, established the principle that users should be free to inspect, modify, and redistribute software. Decades later, a similar crisis is emerging with artificial intelligence, as developers and organizations grow dependent on closed AI models and cloud inference APIs they cannot inspect, reproduce, or control. While open-weight models like Qwen and MiniMax offer partial relief, advocates argue that downloadable weights are not true open source, since the training process remains opaque and outputs cannot be reliably reproduced. The shift toward agentic, cloud-based AI workloads is deepening this dependency, with access to frontier models increasingly restricted to select companies and governments. Critics warn that without meaningful transparency and user control over AI systems, the computing freedoms built over the past half-century are at serious risk of being undone.