Australian government information provides a unique source of evidence about the decisions and activities of government. In this digital age, the volume and complexity of digital information and data created and maintained by government agencies are unprecedented.
The National Archives of Australia (NAA) is responsible for deciding which Australian Government information should be kept or destroyed under the Archives Act 1983. These decisions are enacted through legal documents called records authorities which are issued to agencies. They describe what information is created by agencies and determine how long the information needs to be kept before destruction or retention as national archives.
Presently, National Archives records authorities are developed manually and issued as individual textual documents for manual implementation. They are not suitable for automation without posing a significant risk of destroying valuable information. This limitation places a resource and financial burden on government agencies to accountably manage digital information and data.
The challenge is: How does the National Archives need to change its existing process for identifying which Australian Government information and data must be kept, transferred to the Archives, or destroyed? What format and method will support automatic implementation?
Can we use machine learning, semantic analysis, or predictive coding (or any other sophisticated techniques) to transform the process?
The National Archives is looking for an automated, innovative, accurate and reliable solution to create and manage complex decisions about the value of information and data. Humans can then redirect their efforts towards exceptional and complex decision points. This product would be attractive to governments at all levels, as well as any private sector or not-for-profit organisation that manages information and data.
Existing technology can automatically analyse and identify the content and context of information and data. Sophisticated tools and technologies for automation have improved many previously human-based activities, for example in scheduling in mine operations, disaster management and health (Ref & link: University of Melbourne).
However National Archives research indicates there is no suitable solution that can automatically identify the value of information as described in this challenge. This research comprised national and international environmental scans, investigation of new technologies and their uses, work undertaken with contractors and other vendors, and small in-house tests on internal data.
The solution should:
- deliver a new automated process and tools for issuing National Archives retention and destruction authorisations using advanced technologies. This will replace the current process where National Archives’ experts work with individual agencies and groups of agencies to: analyse government functions and business activities, and the context of individual agencies; identify requirements for creation and management of information and data that documents the work of government, and describe and authorise these requirements in a records authority.
- deliver tools to apply these authorisations in government agencies to automatically identify the value of information, selecting it either for destruction or permanent preservation as evidence of the work and decisions of the Australian Government replacing the current process where agencies manually match or map their information to these requirements
- be based on the most suitable emerging technologies to deliver a digital product that can be tested on National Archives data in a proof of concept
- be consistent with National Archives powers under the Archives Act 1983
- be scalable for use across all types of archives and business sectors and be transferrable across Australian and international jurisdictions.
An expected outcome is to redesign the National Archives records authority creation process using automation, machine learning and artificial intelligence to deliver a digitally produced product. The solution will be used initially to create a records authority for the functions of the National Archives which can be tested on NAA data.
To be eligible you must:
- have an Australian Business Number (ABN)
- and be one of the following entities:
- a company, incorporated in Australia
- an incorporated trustee on behalf of a trust
- an individual or partnership, provided you agree to form a company incorporated in Australia to enter into a grant agreement
- and have a turnover of less than $20 million for each of the last three years. This includes the turnover of any related bodies corporate. If an Australian university or public sector research organisation controls your company, your turnover must be less than $20 million for each of the last three years. Newly established companies are welcome to apply.
You are not eligible to apply if you are:
- an income tax exempt corporation
- an individual or partnership (however, an individual or partnership may apply if they agree to form an eligible corporation before signing a grant agreement)
- a trust (however, a corporation that is a corporate trustee may apply on behalf of a trust)
- a Commonwealth, state or local government agency or body (including government business enterprises).
Terms and Conditions
For complete details on the terms and conditions that govern the Automating complex determinations for Australian Government information challenge please refer to the website.