HOPL: The Human Only Public License (vanderessen.com)

A developer has published the Human Only Public License (HOPL), a permissive-MIT-style license that adds a strict, maximal ban on any use of licensed material by AI systems. HOPL forbids training, fine-tuning, analysis, automated reading, indirect use, and human–AI collaborative workflows that involve the software, its source, outputs or services. Traditional non-AI dev tools (compilers, linters, build systems, debuggers, VCS, static analysis) are explicitly allowed, but AI-powered code completion or ML-based analysis is disallowed. HOPL also includes a copyleft clause requiring derivatives to retain equivalent human-only restrictions, termination on breach, obligations for deployers to publish terms of service forbidding AI access, and a definition of “meaningful human review and creative input.” For the AI/ML community this is significant because it attempts to create a legal barrier to using code, text, or art as training data or as part of ML pipelines. If adopted broadly, HOPL-licensed assets would complicate model training, data collection, scraping, and automated code-assist tools, and could trigger license-scanning alarms in corporate workflows. However, enforceability and detection of AI access remain open questions, and HOPL likely conflicts with open-source norms and existing license ecosystems. The license is publicly posted on GitHub and the author solicits legal feedback, so it may evolve as practitioners and lawyers weigh technical practicality against legal and community ramifications.