Open weights are not open source: Why AI's favorite label is under dispute
Intellectual Property LawComputer ScienceLaw
THE AI ANGLE
Distributing trained neural network weights while keeping training datasets undisclosedA major debate has emerged over whether releasing 'open weights' legitimately qualifies as open source, with critics arguing that withholding training datasets prevents meaningful inspection, reproduction, and accountability. In response, organizations like the Open Source Initiative and the Linux Foundation are grappling with new standards and licensing models, such as OSAID and OpenMDW, to navigate the combination of code, architectures, and weights. For legal and computer science educators, this conflict underscores how machine learning disrupts foundational copyright, licensing, and open-source software definitions.
THE TEACHING ANGLE
Students can examine whether releasing model weights without transparent, auditable training data upholds the legal and ethical standards of open-source licensing or simply amounts to openwashing.Read the original at theregister.com Generate teaching or study materials
More in Intellectual Property Law
- Early Anthropic hire, former METR COO have found a way to rein in rogue AI agentsTechCrunch · September 15, 2026
- AI’s best coding agent fails 60% of the time — and the data backs it upThe New Stack · September 15, 2026
- Ex-FTC boss Khan urges Uncle Sam to break out the handcuffs for AI CEOs, citing 1934 precedentThe Register · September 15, 2026
- Exclusive: Paying for frontier AI models buys 4-month head start at 5x the costArs Technica · September 15, 2026
- RubyGems say OpenAI agents responsible for undisclosed swarm attack against its infrastructureTechRadar · September 15, 2026