Update #82: AI Lawsuits and SOPHON
Evolving US copyright litigation against AI developers may shape training data norms that Australian agencies and vendors will eventually face.
Key points
- US courts are allowing copyright infringement claims against Midjourney, StabilityAI, and Anthropic to proceed.
- SOPHON is a research framework designed to prevent pre-trained AI models from being fine-tuned for harmful tasks.
- Item is a mixed newsletter combining legal commentary and technical research - moderate signal for APS readers.
Summary
This newsletter edition covers two distinct AI topics. First, it surveys the growing wave of US copyright and IP lawsuits against generative AI companies including OpenAI, Midjourney, StabilityAI, and Anthropic, noting courts are beginning to allow fair use and copyright infringement claims to proceed. Second, it introduces SOPHON, an academic framework from Zhejiang University and Ant Group that proposes 'non-fine-tunable learning' to prevent pre-trained models from being repurposed for harmful or unethical tasks. Both threads have governance implications but neither represents an immediate Australian policy development.
Implications for Australian agencies
- Monitor APS policy teams tracking AI regulation may want to monitor US court rulings on fair use and training data, as outcomes could influence Australian copyright and AI governance discussions.
- Monitor AI safety and procurement teams could note SOPHON-style model restriction techniques as an emerging technical control concept for future vendor and model governance assessments.
Implications are AI-generated. Starting points, not advice.
"Update #82: AI Lawsuits and SOPHON" Source: The Gradient – Substack Published: 27 August 2024 URL: https://thegradientpub.substack.com/p/update-82-ai-lawsuits-and-sophon This newsletter edition covers two distinct AI topics. First, it surveys the growing wave of US copyright and IP lawsuits against generative AI companies including OpenAI, Midjourney, StabilityAI, and Anthropic, noting courts are beginning to allow fair use and copyright infringement claims to proceed. Second, it introduces SOPHON, an academic framework from Zhejiang University and Ant Group that proposes 'non-fine-tunable learning' to prevent pre-trained models from being repurposed for harmful or unethical tasks. Both threads have governance implications but neither represents an immediate Australian policy development. Implications for Australian agencies: - [Monitor] APS policy teams tracking AI regulation may want to monitor US court rulings on fair use and training data, as outcomes could influence Australian copyright and AI governance discussions. - [Monitor] AI safety and procurement teams could note SOPHON-style model restriction techniques as an emerging technical control concept for future vendor and model governance assessments. Retrieved from SIMS, 18 May 2026.