"Advanced AI systems are developed and controlled by a small number of private companies," write the authors of this report (74 page PDF). "This concentration of power poses not only economic risks but also significant democratic challenges." I agree. Here's what's needed, they say, to support a public AI alternative (quoted):
- providing strategic public computing resources, particularly supporting open-source AI development;
- creating high-quality datasets as digital public goods through commons-based governance; and
- fostering an ecosystem of fully open source AI models.
I think we would also need the equivalent of an 'open AI practices' section to model and demonstrate how to use AI openly and ethically in the public interest. The report contains a nice overview of AI, with attention to introducing the vocabulary to non-technical readers, contrasts (as illustrated) public and private AI, then maps the roadmap to public AI.
Today: Total: [Direct link] [Share]
