The basic steps to starting a government AI program
​
-
Key initial focus areas for a Government AI program
-
Drafting an AI working program
-
Establishing an AI working group – template guide
-
Developing interim AI guidance
-
Engaging with stakeholders
-
Next steps
​
Key initial focus areas for a government AI program
Drafting an initial AI work program
​
To ensure the government is capable of maximising the opportunities of AI and managing the risks. Outline priority short, medium, and long-term activities.
Short-term
Develop interim guidance for government use of generative AI tools
-
Assess and expand upon the risk classification and assessment in the interim guidance.
Develop approach on governance
-
Develop an approach to AI governance that provides a more consistent and fit for purpose model for agencies.
​​
Investigate feasibility of a policy for the responsible use of AI in government
-
Investigate the feasibility of policy development that covers government-specific AI issues and the rules that guide the responsible use of AI in agencies.
​
Medium-term
​
Develop approach on risk management and guidance approach that would help agencies assess AI risk and apply appropriate mitigations.
-
Develop a risk management approach and provide guidance on how to use it.
Develop policies relating to the responsible use of AI
-
Develop a policy for the responsible use of AI in government if it is found to be feasible.
​
Establish AI Proof of Concepts
-
Drawing on the expertise of leading agencies already using AI, establish Proof of Concepts (PoC) and /or build on existing use cases across a range of use case categories.
​​
Develop technical approaches for rolling out AI tools in government
-
Work with leading agencies to establish case studies for how various categories of technical use cases can be designed, developed and deployed in agencies.
Develop approach on preparedness
-
Work with relevant national security and intelligence agencies to prepare a series of patterns / processes for agencies to follow when relevant circumstances arise.
​
Long-term
​
Develop a Data and Digital Workforce Plan to address capability gaps
-
Develop plan and ongoing monitoring and evaluation of government capabilities.
Key messages for interim guidance for AI use
​
Note that generative AI offers innovative opportunities for government
-
Due to its rapid evolution and ease of access, there is an urgent need and growing demand for guidance
-
Interim Government guidance for government agencies has been developed for users
-
This is just the start and more guidance will be developed over the coming months
​
Taking a principles-based approach for AI in government
​
-
Ensure AI use cases are safe, responsible, and ethical
-
Deploy AI responsibly
-
Seek to be transparent and provide explainability
-
Embed privacy and security
-
Use human centred decision making
​
Provide some tactical guidance for implementing the principles and note more will follow
When using generative AI platforms, you should:
-
only use your corporate credentials to sign up
-
obtain approval before using
-
review all generated content and consider links and files malicious until vetted
-
report instances where you can’t apply the guidance
Next Steps
​
Begin work on the core areas:
-
Governance
-
Risk management
-
Skills and capability
-
Technical
-
Preparedness