The Data Collaboration Framework is a pioneering national standard that redefines the relationship between people and AI systems. Spearheaded by Ask AI Founder Chris McLellan, it's all about reducing copies in order to make true data ownership, control, and collaboration possible.
Last updated: December 2024
A new standard in data, AI, and innovation
Developed through the Digital Governance Council and ratified by the Standards Council of Canada, the Data Collaboraiton Framework was co-authored by Ask AI Founder Chris McLellan, who also serves as Technical Committee Leader.
Currently under review by the International Standards Committee for international adoption, it acts as a "data-level" foundation for a more equitable, responsible, and safe AI industry.
Protecting sensitive data like property
We don't think about it often, but the fact is that societies around the globe protect things of value, including property, currency, and IP, by making them difficult (and illegal) to copy . But when it comes to personal and senstitive data, it's basically been a free-for-all for decades.
In order to create a more equitable AI industry, sensitive and personal data needs to be protected more like physical property
The fact is that traditional, copy-based data integration (i.e. sharing data beween apps and systems) makes it extremely difficult—if not impossible, to meaningfully control and protect data.
The Data Collaboration Framework takes a "Gordian Knot" approach this complex issue by encouraging developers to take a more data-centric approach to innovation, where a single uncopied data source can be used to power unlimited digital solutions.
By eliminating copies, data owners can set access controls once (as metadata), avoiding the need for developers to manage the controls (in code) across countless applications—where they become inconsistent or ignored altogether.
Data Collaboration Framework: Core Principles
Traditional digital innovation relies on integrating copies of data between individual applications which is antithetical to control, governance, and compliance.
The Data Collaboration Framework turns this situation on its head by giving data owners the power to grant access, not copies, to individual developers and solutions.
Key principles:
Decouple data from apps (shared architectures, not silos)
Establish data products, decentralize governance
Prioritize active metadata over complex code
Access-based collaboration, not copy-based integration
Set access controls at metadata-layer, not application code
Prioritize composability over monolithic solutions
By integrating these principles, the Data Collaboration Framework enables more responsible AI by improving efficiency while ensuring that humans retain control over the data they contribute. This has obvious and immediate implications for the training of AI models, and while it's not a panacea (can data be revoked from existing AI models?) it represents a significant step towards HUMAN CONTROL of the raw materials used by AI systems to deliver value.
Data Collaboration Framework timeline
2025 - Review by ISO for pontential adoption
July 2024 – Updates published by Technical Committee
February 2023 – Ratification by Standards Council of Canada
2022 – Drafting and peer reviews at Digital Governance Council
Links & Resources
Download the Standard – Access the full framework
Standards Council of Canada – Learn about national standard
ISO/IEC JTC 1/SC 32 – Learn about the international review process
Track our progress
Join the Ask AI newsletter if your're interested in following our efforts to help people spot the opportunities and minimize the risks associated with the increased adoption of AI at work.
Comments