AI, Copyright, and Academic Research Production
In a significant policy decision, Australia’s Attorney-General Michelle Rowland has ruled out introducing a text and data mining (TDM) exception to the country’s copyright law—a move that would have permitted artificial intelligence (AI) developers to freely scrape and use copyrighted creative works for AI training without compensating rights holders. This proposal, initially floated by the Productivity Commission in its August interim report, was met with strong opposition from creatives, authors, journalists, and cultural institutions, who argued it would amount to the “theft” of intellectual and cultural property by big tech firms. The government’s rejection of the TDM exception has been widely celebrated as a victory for creators and a reaffirmation of the value of Australia’s cultural output. While the final Productivity Commission report—due in December—may still include the recommendation, the immediate decision, coupled with the reactivation of the Copyright & AI Reference Group (CAIRG) in late October, signals a shift toward a licensing-led framework for AI development. This approach emphasizes negotiated, consensual use of copyrighted materials rather than blanket exemptions. For the academic community, this decision carries both reassurance and complexity. On one hand, it upholds strong copyright protections that safeguard scholarly output, teaching materials, and research publications—many of which are protected under the same legal regime as creative works. This is especially important in an era where AI systems increasingly ingest academic content without attribution or compensation, potentially undermining the economic and moral rights of researchers and institutions. On the other hand, academics engaged in AI research or computational […]