Championing open-source and ethical AI for all
Join the movement for transparency, fairness, and innovation in artificial intelligence. Together, we can break the chains of proprietary AI and build a future where technology serves humanity.
Join the MovementWe believe artificial intelligence should be open, ethical, and accessible to everyone. Our mission is to advocate for AI systems that prioritize human values and democratic principles.
Open-source AI allows anyone to inspect, understand, and verify how systems make decisions. No black boxes, no hidden agendas—just clear, accountable technology that earns trust through openness.
Ethical AI development means addressing bias, ensuring equitable access, and designing systems that respect human rights. We advocate for AI that serves all communities, not just the privileged few.
When AI is open, innovation flourishes. Researchers, developers, and creators worldwide can build upon shared knowledge, accelerating breakthroughs that benefit humanity rather than corporate bottom lines.
Four foundational principles guide everything we do at Liberate AI.
We champion AI models, datasets, and tools that are freely available for study, modification, and distribution. Open source is the foundation of democratic AI development.
Responsible AI requires clear ethical frameworks and community-driven governance. We advocate for policies that prioritize safety, accountability, and human dignity in AI development.
AI belongs to everyone. We build and support inclusive communities where developers, researchers, activists, and citizens collaborate to shape the future of technology together.
Rigorous, reproducible research is essential for advancing AI responsibly. We promote open access to research findings and support initiatives that preserve scientific knowledge for future generations.
Preserving AI history is crucial for research, education, and accountability. Our digital sanctuary archives legacy models so they remain accessible for study and reproducibility.
As AI evolves rapidly, older models often disappear when companies pivot or shut down. These models represent important milestones in AI development and contain valuable insights for researchers studying the evolution of machine learning.
Our sanctuary ensures that historic AI models remain available for:
OpenAI's GPT-2 was a landmark language model that demonstrated impressive text generation capabilities. Initially withheld due to safety concerns, it was later released and became a cornerstone for open-source language model research.
Google's Bidirectional Encoder Representations from Transformers revolutionized natural language understanding. BERT's pre-training approach influenced countless subsequent models and remains essential for NLP education.
Microsoft's Residual Networks introduced skip connections that enabled training of much deeper neural networks. ResNet architectures remain foundational in computer vision and continue to be widely used and studied.
Google's Word2Vec introduced efficient methods for learning word embeddings from large text corpora. This model laid the groundwork for modern language understanding and remains a key teaching tool in NLP courses worldwide.
Support our sanctuary by contributing computing resources, donating to cover hosting costs, or helping us locate and archive legacy models.
Join us in the fight for open, ethical AI. There are many ways to contribute to the movement.
Stay informed about our campaigns, events, and the latest news in open-source AI.
Help build tools, contribute to our sanctuary, or improve our documentation. Every contribution matters.
View on GitHubSupport our advocacy work, sanctuary hosting costs, and community programs with a financial contribution.
Join our workshops, webinars, and conferences to connect with fellow advocates and learn about ethical AI.
Stay up to date with our latest campaigns, research, and community updates.
We're excited to announce a collaboration with leading universities to promote open-source AI research and education. This partnership will provide students with access to legacy models and research tools.
Read MoreOur latest policy brief outlines why governments should require transparency in AI systems used for public services. We present evidence from case studies and propose actionable legislation.
Read MoreA milestone for AI preservation! Our sanctuary now hosts 50 historic AI models, from early neural networks to pioneering language models. Thanks to our community for making this possible.
Read More