Artificial intelligence transparency statement

Artificial Intelligence (AI) is transforming the way we work, making tasks faster and more efficient. We recognise there are risks and we are committed to using AI appropriately – always following applicable law and meeting the public’s expectations.

You can expect that we will:
• ensure all AI use will meet privacy, security and legal requirements 
• only use AI tools that have been properly assessed for accuracy and fairness
• only use AI systems designed to avoid bias and that are accessible for everyone
• always use AI with human oversight
• train our staff to use approved tools and keep improving how we use AI.

The use of AI in elections

AI is never used to run elections in Victoria or determine results. Our use of AI is limited to administrative functions that have no impact on how votes are counted. This ensures that the integrity, security, and fairness of every election remain entirely in human hands.

Our use of AI 

We are committed to using AI responsibly, with strict safeguards, human oversight and transparency.

We will follow privacy and security laws

Every AI tool we use will comply with privacy, security, and data protection rules. We will regularly review our use of AI to ensure we continue to meet these requirements.

Clear accountability

The Electoral Commissioner is the accountable officer responsible for AI use in our organisation. The Chief Information Officer will oversee how AI is managed routinely.

Managing risks

We recognise that AI comes with risks. We will carefully identify and manage these risks in accordance with our risk management framework.

Assessment before using

No AI tool will be put into use until it has been assessed for accuracy, fairness, and alignment with our standards.

Checking for bias

AI systems will be regularly reviewed to ensure they do not generate biased or unfair outputs. Issues will be addressed as they arise.

Fair and accessible design

Internally developed AI tools will be designed to be fair, inclusive, and accessible to all users.

Respecting copyright and intellectual property

We’ll only use AI systems, datasets, and content that are properly licensed. We will not employ copyrighted or proprietary materials for AI training unless explicit authorisation has been secured from the rights holder.

Measuring performance

We will monitor the effectiveness of our AI tools by assessing factors such as accuracy, response times, and user satisfaction. Performance will be regularly evaluated and subject to ongoing improvement initiatives.

Training our staff

Staff in the Victorian Electoral Commission will be trained so they understand the risks of AI, and how to use it responsibly.

Human oversight

Humans will always be involved to ensure AI is used responsibly and ethically and to maintain accuracy and integrity.