Australia - Tasmania - AI Use Guidelines (2024)

Overview

The 'Guidance for the use of artificial intelligence in Tasmanian Government' serves as a pivotal document for the Tasmanian public service, establishing a clear and consistent framework for the ethical, safe, and responsible adoption of Artificial Intelligence (AI) technologies across all government agencies. Approved by the Tasmanian Government Secretaries Board on September 13, 2024, this guideline acknowledges the dual nature of AI as both a significant driver of opportunity and a source of potential risks. Its primary objective is to ensure that the Tasmanian Government can harness the transformative potential of AI to enhance efficiency, boost productivity, streamline processes, improve service design and delivery, and refine policy development through advanced data analysis, while simultaneously safeguarding public trust and upholding core values. The document is explicitly designed to be dynamic, recognizing the rapid evolution of AI technologies and committing to periodic updates to reflect new opportunities, emerging risks, and technical advancements, thereby ensuring its continued relevance and effectiveness in a fast-changing digital landscape.

This guidance is not merely a set of prescriptive rules but a strategic enabler, encouraging innovation within a controlled and accountable environment. It underscores the Tasmanian Government's commitment to proactive engagement with advanced technologies, positioning the state to leverage AI for long-term prosperity and improved public outcomes. By providing a baseline approach, the guidance aims to prevent fragmented or inconsistent AI deployments across different agencies, fostering a unified and coherent strategy for AI integration. It aligns closely with broader national initiatives, including the development of a national framework for AI assurance in government, demonstrating a commitment to a harmonized approach to AI governance across Australian jurisdictions. This overarching alignment ensures that Tasmanian agencies benefit from shared knowledge and best practices, contributing to a robust and trustworthy AI ecosystem.

Definitions

To ensure clarity and consistency across all government agencies engaging with AI, the guidance provides fundamental definitions of key terms. Artificial Intelligence (AI) is broadly defined as a domain of computer science focused on creating systems that can emulate human behaviour, particularly in learning and autonomously executing actions. More specifically, it refers to engineered systems capable of generating predictive outputs such as content, forecasts, recommendations, or decisions. This foundational definition helps to delineate the scope of technologies covered by the guidance, ensuring that agencies understand what constitutes an AI system in the context of government operations and service delivery. The emphasis on 'engineered systems' and 'predictive outputs' highlights the practical application and impact of AI within the public sector, from automated administrative tasks to data-driven policy insights.

A critical distinction is made for 'Generative AI,' which is characterized as a subset of AI capable of producing novel content, including text, images, audio, and video, that closely resembles human-generated output. Generative AI excels at recognizing patterns and replicating them when tasked with creation. Examples like Microsoft Copilot are cited as tools with the potential to significantly assist employees by automating routine document drafting and other repetitive tasks, thereby freeing up human resources for more complex work. These definitions are crucial for fostering a shared understanding across government, enabling consistent application of the guidance's principles and recommendations. Without clear definitions, there is a risk of misinterpretation or inconsistent deployment, which could undermine the objectives of safety, ethics, and trustworthiness that underpin the entire framework.

Governance and Institutional Framework

The governance of AI within the Tasmanian Government is structured to ensure a consistent, responsible, and accountable approach across all agencies. The 'Guidance for the use of artificial intelligence in Tasmanian Government' was formally approved by the Tasmanian Government Secretaries Board on September 13, 2024, signifying a high-level commitment and endorsement from the state's senior public service leadership. This approval mechanism ensures that the guidance carries significant weight and is intended to be implemented uniformly throughout the government. The Digital Strategy and Services unit, likely within the Department of Premier and Cabinet, plays a crucial role in developing, maintaining, and contributing to a range of policies, strategies, guidelines, and toolkits that support consistency in the management and use of digital data and services, including those related to AI. This centralized oversight ensures strategic alignment and facilitates knowledge sharing across the public sector.

While the guidance establishes a whole-of-government baseline, it also mandates that individual agencies develop their own specific policies or guidance. These agency-specific frameworks must align with the broader government and industry standards and frameworks, ensuring tailored application while maintaining overall consistency. This federated approach to governance recognizes that different agencies will have unique operational contexts and varying levels of AI adoption, requiring bespoke policies that address their specific needs and risks. The emphasis on whole-of-government cooperation and collaboration is a cornerstone of this framework, promoting a culture of shared learning and resource optimization. This collaborative ethos extends to working with Commonwealth and other jurisdictions to develop a nationally consistent approach, reducing duplication of effort and ensuring that Tasmania benefits from national best practices in AI governance.

Key Focus Areas

The Tasmanian AI Guidance for Government outlines seven core recommendations that serve as the key focus areas for agencies engaging with artificial intelligence. Firstly, agencies are urged to deploy AI responsibly, ensuring that all AI systems are implemented in a manner that is safe, trustworthy, and ethical. This foundational principle underpins all subsequent recommendations, emphasizing the paramount importance of public confidence and adherence to ethical standards. Secondly, the guidance advocates for the adoption of a robust risk-based approach for specific uses and applications of AI. This involves systematically assessing the potential impact and risks associated with each AI deployment and tailoring mitigation strategies accordingly, aligning with the agency's defined risk appetite. This proactive risk management strategy is crucial for navigating the complexities and uncertainties inherent in AI technologies.

Further key areas include the development of agency-specific policies or guidance that are consistent with overarching government and industry standards, fostering a tailored yet harmonized approach to AI governance. A commitment to adopting a consistent whole-of-government vocabulary for AI is also emphasized, ensuring clear communication and understanding across the public service. Building awareness and capabilities among staff to effectively develop, deploy, and operate AI systems forms another critical focus, recognizing that human expertise and oversight are indispensable for successful AI integration. The guidance also stresses the importance of aligning procurement practices with responsible deployment and risk assurance processes for AI, ensuring that external AI solutions meet the government's ethical and safety standards. Finally, a strong commitment to whole-of-government cooperation and collaboration is encouraged, promoting knowledge sharing and collective problem-solving in the evolving AI landscape. Beyond these recommendations, the guidance also highlights critical considerations for privacy and data security, mandating compliance with legislation such as the Personal Information Protection Act 2004, and robust protective and cyber security practices to safeguard sensitive information used within AI solutions.

Implementation Framework

The implementation framework for the Tasmanian AI Guidance for Government is designed to integrate AI responsibly and effectively across the public service, building upon existing legislative and policy foundations. A core tenet of this framework is the explicit alignment with national initiatives, particularly the 'National framework for the assurance of artificial intelligence in government' and Australia's AI Ethics Principles. This commitment to national consistency ensures that Tasmanian agencies benefit from a broader, harmonized approach to AI governance, reducing fragmentation and leveraging shared expertise. The guidance encourages agencies to use these national frameworks as a blueprint for their internal policies and practices, adapting them to the specific context of Tasmanian government operations. This alignment also facilitates cross-jurisdictional collaboration and ensures that AI deployments in Tasmania are subject to a similar standard of risk assessment and ethical consideration as those in other Australian states and territories.

Agencies are expected to translate the high-level recommendations into actionable internal policies and procedures. This includes developing clear guidelines for the responsible deployment of AI, establishing internal risk assessment processes, and fostering a culture of continuous learning and capability building among staff. The framework also emphasizes the importance of integrating AI considerations into existing procurement processes, ensuring that any AI solutions acquired from external vendors adhere to the government's ethical, safety, and security standards. Furthermore, the guidance implicitly calls for a proactive approach to data governance, requiring agencies to ensure that data inputs into AI systems, especially public or 'open' AI tools, do not compromise sensitive, classified, or personal information. Compliance with the Personal Information Protection Act 2004 is explicitly highlighted as a critical legal requirement, underscoring the importance of privacy and data security throughout the AI lifecycle.

Monitoring and Evaluation

The 'Guidance for the use of artificial intelligence in Tasmanian Government' acknowledges the dynamic nature of AI technologies and the need for continuous adaptation, stating that the guidance "will be updated periodically, to reflect new developments." This commitment to periodic review forms the cornerstone of its monitoring and evaluation framework. While the document does not explicitly detail a formal, structured monitoring and evaluation program with specific metrics or reporting cycles, the intent is clear: the Tasmanian Government recognizes that AI is an evolving field, and its guidance must remain agile and responsive to emerging opportunities, identified risks, and technical advancements. This implies an ongoing process of assessment, where the effectiveness of the guidance in promoting safe, trustworthy, and ethical AI deployment is continuously observed and evaluated by relevant government bodies, such as the Digital Strategy and Services unit and the Secretaries Board.

The risk-based approach advocated by the guidance inherently incorporates elements of monitoring and evaluation at the agency level. By requiring agencies to assess the risk of AI impacts in specific contexts, there is an implicit expectation that these assessments will be revisited and updated as AI systems evolve or as new risks become apparent. This continuous assessment at the point of use contributes to an overall understanding of the guidance's efficacy. Furthermore, the emphasis on whole-of-government cooperation and collaboration suggests that insights and lessons learned from AI deployments across different agencies will be shared, contributing to a collective monitoring effort. This collaborative intelligence gathering would inform future updates to the guidance, ensuring it remains relevant and effective in fostering responsible AI innovation while mitigating potential harms. The alignment with national frameworks also means that Tasmania's approach to AI will be implicitly evaluated against broader Australian standards and evolving international best practices.

Penalties, Liability, and Appeals

As a non-binding guideline, the 'Guidance for the use of artificial intelligence in Tasmanian Government' does not directly prescribe specific penalties, establish new liability regimes, or outline formal appeal processes for AI-related issues. Its primary function is to provide recommendations and best practices for responsible AI deployment rather than to create new legal obligations. However, the guidance explicitly reinforces the necessity for agencies to comply with existing legal frameworks, which inherently carry provisions for penalties and liability. For instance, it mandates that AI solutions and initiatives must be compliant with the Personal Information Protection Act 2004. Any breaches of this Act, particularly concerning the handling of personal information by AI systems, would fall under the existing enforcement mechanisms and potential penalties outlined in that legislation. Similarly, adherence to protective and cyber security standards for Tasmanian Government agencies is required, meaning failures in these areas due to AI deployments would be subject to existing government policies and potentially broader legal consequences related to data breaches or security incidents.

The emphasis on responsible deployment, ethical considerations, and a risk-based approach within the guidance serves as a proactive measure to prevent situations that might lead to legal challenges or public complaints. By advising agencies to ensure human oversight and accountability for AI outcomes, the guidance implicitly directs responsibility to human decision-makers within the government structure, rather than attributing it solely to autonomous systems. In cases where AI-driven decisions lead to adverse outcomes for individuals, existing administrative law mechanisms for review and appeal of government decisions would likely apply. While the guidance itself does not create these pathways, it operates within a broader legal and administrative landscape where citizens have rights to redress. The ongoing development of national AI frameworks and potential future legislation may introduce more specific provisions for AI liability and appeals, which the Tasmanian guidance would then need to integrate through its periodic updates.

Relationship to Other Instruments

The 'Guidance for the use of artificial intelligence in Tasmanian Government' is not a standalone document but is strategically positioned within a broader ecosystem of national and state-level legislative and policy instruments. A cornerstone of its framework is the explicit alignment with the 'National framework for the assurance of artificial intelligence in government.' This national framework, developed through collaborative efforts between the Commonwealth and other Australian jurisdictions, aims to establish a consistent and ethical approach to AI use across all levels of government. By aligning with this national standard, Tasmania ensures that its AI deployments are harmonized with broader Australian practices, benefiting from shared expertise and reducing the potential for regulatory inconsistencies. The Tasmanian guidance also incorporates and aligns with Australia's AI Ethics Principles, which provide a set of foundational values for responsible AI development and use, further embedding ethical considerations into the state's approach.

Beyond national AI-specific frameworks, the guidance also integrates with existing Tasmanian legislation and government policies. Critically, it mandates that all AI solutions and initiatives comply with the Personal Information Protection Act 2004. This ensures that the deployment of AI systems respects individual privacy rights and adheres to established standards for the collection, use, storage, and disclosure of personal information. Furthermore, the guidance requires adherence to the protective security standards applicable to Tasmanian Government agencies and their subsidiaries, including specific requirements for information management and cyber security. This integration ensures that AI systems are developed and operated within a secure digital environment, mitigating risks associated with data breaches and cyber threats. The guidance also complements other Tasmanian Government policies related to digital strategy, cloud computing, and information management, forming a cohesive policy landscape for digital transformation within the state.

International Alignment

While the 'Guidance for the use of artificial intelligence in Tasmanian Government' is a state-level document within Australia, its underlying principles and strategic alignments demonstrate a strong commitment to internationally recognized best practices in AI governance. The guidance explicitly states its alignment with the 'National framework for the assurance of artificial intelligence in government,' which itself is developed in collaboration with Commonwealth and other jurisdictions to provide a nationally consistent approach to safe and ethical AI use. This national framework, in turn, aligns with Australia's AI Ethics Principles. These principles, while national, are broadly consistent with evolving international ethical AI guidelines and frameworks, such as those promoted by the OECD, UNESCO, and the European Union. By adhering to these nationally recognized principles, Tasmania indirectly benefits from and contributes to a global consensus on responsible AI development and deployment.

The emphasis on ethical AI, human oversight, transparency, accountability, and a risk-based approach resonates deeply with international dialogues and recommendations on AI governance. Many international bodies advocate for similar foundational principles to ensure AI systems are developed and used in ways that respect human rights, promote fairness, and minimize societal risks. The Tasmanian Government's proactive stance in developing this guidance, and its commitment to periodic updates, reflects an understanding that AI governance is a global challenge requiring continuous learning and adaptation from international developments. Although the document does not directly reference specific international treaties or agreements, its foundational alignment with national frameworks that are themselves informed by global best practices ensures that Tasmania's approach to AI is compatible with, and contributes to, the broader international effort to govern AI responsibly.

Implementation Timeline

Sources and References