Update #82: AI Lawsuits and SOPHON
A broad AI landscape digest surfaces ongoing copyright litigation and model-misuse research threads worth monitoring for risk and governance implications.
Key points
- A multi-topic AI newsletter covering copyright lawsuits, a novel misuse-prevention technique, and AI governance vignettes.
- The SOPHON research introduces a framework to prevent pre-trained models being fine-tuned for harmful or restricted tasks.
- Primarily US-focused content with limited direct APS relevance; useful as a broad AI landscape signal.
Implications for Australian agencies
- Monitor Risk and governance teams may want to monitor US court rulings on AI copyright and fair use, as outcomes could influence how Australian agencies assess IP risk in AI procurement and deployment.
- Monitor Researchers and policy teams tracking AI misuse controls may want to note the SOPHON framework as an emerging technical approach to restricting harmful fine-tuning of released models.
Implications are AI-generated. Starting points, not advice — see methodology for how they're framed.
View original source
Copied.
"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
The Gradient's Update #82 is a curated newsletter covering several distinct AI developments from August 2024. The lead items address the growing wave of US copyright and IP litigation against AI companies including OpenAI, Midjourney, Stability AI, and Anthropic, with courts beginning to allow fair-use and direct-infringement claims to proceed. A research summary covers SOPHON, a Zhejiang University and Ant Group framework using inverse meta-learning to make pre-trained models resistant to fine-tuning for harmful or restricted tasks. Shorter items cover an AI-run mayoral campaign in Wyoming, Amazon's comments on coding automation, DeepMind worker protests over defence contracts, open-source AI definitions, and an arrest for AI-generated child sexual abuse material.
Implications for Australian agencies:
- [Monitor] Risk and governance teams may want to monitor US court rulings on AI copyright and fair use, as outcomes could influence how Australian agencies assess IP risk in AI procurement and deployment.
- [Monitor] Researchers and policy teams tracking AI misuse controls may want to note the SOPHON framework as an emerging technical approach to restricting harmful fine-tuning of released models.
Retrieved from SIMS, 18 July 2026.