Description

Large language models would be subject to high-risk regulation and tech companies would need to pay artists for training the technology under an EU-style whole of economy AI Act proposed by the Senate. The Senate’s Select Committee on Adopting Artificial Intelligence, which struggled to get answers from companies like Meta and Google, made the recommendations on Tuesday night as the government grapples with its regulatory approach. It also recommended the government do more to support the local AI industry, but stopped short of endorsing calls for direct involvement in a sovereign general purpose AI model.

Summary

The Australian Senate has proposed a comprehensive AI Act that mirrors the European Union's regulatory framework, with a focus on high-risk regulations for large language models and fair compensation for artists whose work is used in AI training. This proposal emerges amidst ongoing challenges in obtaining satisfactory responses from major tech companies such as Meta and Google. While the Senate's Select Committee on Adopting Artificial Intelligence suggests measures to bolster the local AI sector, they have not advocated for direct government development of a sovereign AI model. The proposed legislation could significantly influence Australia’s approach to managing AI-related risks and ensuring ethical AI development.

Body