The paper further investigates the history of open source, its separation from free software, and its adoption and manipulation by large tech corporations. It uses OpenAI as an example of the overbroad and ill-defined use of the term 'open' by tech companies. The authors argue that while 'open' AI can provide transparency, reusability, and extensibility, it does not ensure democratic access or meaningful competition in AI, nor does it solve the problem of oversight and scrutiny. They conclude that some companies use 'open' AI to bolster their positions and entrench dominance, using the rhetoric of 'open' AI to expand market power while benefiting from the free labor of open source contributors.
Key takeaways:
- The terms 'open' and 'open source' are often used in confusing and diverse ways in the context of AI, often blending concepts from both open source software and open science.
- While some maximally open AI systems exist, the resources needed to build AI from scratch and to deploy large AI systems at scale remain 'closed' and available only to those with significant resources.
- The term 'open' AI is being used by tech companies to shape the public's and policymakers' understanding about AI, its capabilities, and the power of the AI industry.
- Some companies have moved to embrace 'open' AI as a mechanism to entrench dominance, using the rhetoric of 'open' AI to expand market power while benefiting from the free labor of open source contributors.