Artificial Intelligence (AI) is a revolutionary technology. Compared to the unprecedented innovations of the past two decades, AI stands out as a unique and pivotal development that has the potential to impact people worldwide. We increasingly recognize its ability to expedite pharmaceutical drug development, enhance energy efficiency, transform cybersecurity, and improve accessibility.
As we continue to broaden the scope of AI applications and make technological advancements, it has become increasingly essential to ensure that our efforts are not isolated. It is crucial for industry, governments, researchers, and civil society to join forces in bold and responsible collaboration. This collective effort allows for the expansion and sharing of knowledge, the identification of methods to mitigate emerging risks and prevent misuse, as well as the further development of tools to enhance content transparency for people worldwide.
This approach has been ingrained in our work from the outset. Today, we are pleased to share some of the partnerships, commitments, and codes through which we are actively engaged in to comprehend AI’s potential and shape its progression responsibly.
Collaborations, Partnerships, and Frameworks with Industry
- Frontier Model Forum: Google, alongside Anthropic, Microsoft, and OpenAI, launched the Frontier Model Forum to further the safe and responsible development of advanced AI models. The Forum, in association with philanthropic partners, has pledged over $10 million for a new AI Safety Fund to propel research into tools for society to effectively test and evaluate the most powerful AI models.
- Partnership on AI (PAI): We played a role in the development of PAI, which functions as a community of experts committed to nurturing responsible practices in the development, creation, and dissemination of AI, including media generated using AI technologies.
- MLCommons: As a participant in MLCommons, we contribute to a collective effort to accelerate machine learning innovation and amplify its positive impact on society.
- Secure AI Framework (SAIF): We have introduced a framework for secure AI systems to address risks specific to AI systems, such as theft of model weights, contamination of training data, and introduction of malicious inputs through prompt injection. Our objective is to collaborate with industry partners to implement this framework over time.
- Coalition for Content Provenance and Authenticity (C2PA): We have recently joined the C2PA as a steering committee member. This coalition is a cross-industry initiative aimed at enhancing transparency and providing context for digital content consumers. We are committed to contributing to the development of its technical standard and promoting the adoption of Content Credentials, which consist of tamper-resistant metadata that demonstrates the creation and editing history of content.
Engagement with Governments and Civil Society
- Voluntary White House AI commitments: In alignment with other companies at the White House, we have jointly committed to advancing responsible practices in the development and utilization of artificial intelligence to ensure that AI benefits everyone. Our significant progress in fulfilling our commitments demonstrates our dedication.
- G7 Code of Conduct: We demonstrate our support for the G7’s voluntary Code of Conduct, which is designed to promote safe, reliable, and secure AI worldwide.
- US AI Safety Institute Consortium: We are actively involved in NIST’s AI Safety Institute Consortium, where we share our expertise as part of our collective effort to advance safe and trustworthy AI on a global scale.
- UK AI Safety Institute: The UK AI Safety Institute has access to some of our most advanced models for research and safety purposes, with the aim of building expertise and capability for the long term. Together, we are actively working towards establishing more rigorous evaluations for AI models and seeking consensus on best practices for the evolving sector.
- National AI Research Resource (NAIRR) pilot: We are contributing our cutting-edge tools, computing power, and data resources to the National Science Foundation’s NAIRR pilot, which seeks to democratize AI research across the U.S.
As we continue to expand these efforts, we will consistently update this list to reflect the latest collaborative initiatives with the industry, governments, civil society, and other stakeholders.
