A new high-level AI Governance Code has been launched by the Medical Software Industry Association (MSIA) and Medical Technology Association of Australia (MTAA).

It sets AI governance quality standards for organisations that use, develop, design, manufacture, sponsor, distribute, deploy and/or supply AI Systems (for use) in a Health Context, such as AI-enabled Health Software.

The MSIA and MTAA said the AI Governance Code would assist the health community to navigate its use of unregulated AI software.

The two organisations said the code had been kept deliberately high-level and was accompanied by an AI Governance Accreditation Standard.

The Code acknowledges the Australian Government’s AI Ethics Principles as an important reference point for AI governance and draws on those principles in its own General Ethical Principles.

The Principles expect each participant organisation to use, deploy, develop, design, manufacture, sponsor, distribute, implement and/or supply AI Systems (including Health Software that incorporates an AI System) in a way that “considers impacts and benefits on the individuals, society, and the environment, including by prioritising patient safety and the enhancement of patient care.”

Signatories are also expected to be respectful of human rights, dignity, and the autonomy of individuals and to be fair, inclusive, tested for, and equipped with, proportionate and appropriate bias and discrimination control measures.

Other expectations are for organisations to use AI in a way that aligns with applicable privacy laws in the relevant jurisdiction and protects data security” and with “appropriate systems, policies and processes for testing for safety, reliability, and operation (in accordance with intended purpose) in place.

The Code also expects signatories to engage AI in a way that “promotes transparency and responsible disclosure to support consumers to understand when and how AI Systems engage with and impact them,” as well as “in an accountable way, ensuring that those responsible for AI Systems are identifiable, and that appropriate human oversight is maintained.”

The MSIA and MTAA said they had been working with their members to have the Code ready to fit with the National AI Plan which was announced on December 2.

“We look forward to the health and productivity benefits that the safe use of AI in healthcare will provide,” the MSIA said in a statement.