With the accelerating pace of technology and the rapid growth of AI, the need to assure the trustworthiness of an organisation’s data practices is urgent. The UK government’s AI Safety Summit highlighted many opportunities and threats brought about by new technology, demonstrating its importance globally.
But to have trust in the widespread adoption of AI, consideration needs to be given to data and the data practices followed by those developing and using the technology. AI products and services are built using data, and understanding the ethics, opportunities and limitations is essential.
The Open Data Institute (ODI) has leveraged its extensive experience, research and commitment to accessibility to develop a data practices framework that has the potential to benefit organisations, drive economic growth and promote confidence in data sharing and use.
The framework of nine Data Practices helps people to understand - and then fill the gap - between strategic and technical capabilities within organisations.
The nine practices cover the areas of accountability, privacy, security, standardisation, resourcing, capability, engagement, ethics and permissions. They are the foundational thinking behind a groundbreaking new set of tools and guidelines in development by the ODI.
Research shows a significant gap in the market for easily understandable guidelines, and meeting this untapped potential could deliver 60% growth in the global data assurance sector within four years, according to market research.
In the same way that Cyber Essentials is a United Kingdom certification scheme designed to show an organisation has a minimum level of protection in cyber security, there is a need for organisations responsible for data to be able to measure and demonstrate to all of their stakeholders the trustworthiness of their data practices.
Trustworthy AI depends on organisations developing and implementing AI using the underpinning data in a trustworthy and transparent manner. Although many initiatives are looking at 'data assurance, ' most of the work is done on technical implementation and measurements at a dataset level. The ODI believes there is a need for guidance and support accessible to decision-makers within organisations who must understand and consider what is required to develop trust in their data practices.
By filling this gap in the market, the ODI will eventually pave the way for a system for accreditation and certification. This will enable organisations responsible for data to demonstrate the trustworthiness of their data practices.
Based on the framework, the first product in development is the Data Practices Assessment Tool, launched for beta test applications at the ODI Summit on 7th November. The tool uses the principles of the nine practices to create the first interactive, practical system to help organisations and teams understand and communicate the strengths and weaknesses in their data practices.
Also launched at the ODI Summit was the Data Sharing Risk Assessment, enabling organisations responsible for data to prove the trustworthiness of their data practices and demonstrate that they take understanding the risks of collecting, using and sharing data seriously.
There is a real and urgent need to help organisations understand what it means to be trustworthy with data. Our new Data Practices will help organisations bridge the gap between technical implementations and organisational approaches to engender trust, which everyone can adopt.
Principle Data Specialist, the ODI
Speaking before the ODI Summit 2023, Andrew continued: “We are developing the Data Practice Assessment Tool and the Data Sharing Risk Assessment to enable organisations responsible for data to audit and prove the trustworthiness of their data practices.
“We are committed to creating products and services that work for everyone, so we invite people to join a private test group for the tools at the ODI Summit. This is so that we can take into account feedback from the people who will one day be the end users of the product and ensure that it is clear, accessible and fits the purpose of helping organisations confidently develop data practices that are trustworthy and trusted.”
You can also contact the team directly at [email protected]