This week I had the pleasure of being part of two panel sessions focused on AI governance for Information Awareness Week, a joint collaborative between organisations in the records, archives, library, knowledge, information, and data management communities (National Archives of Australia; Australian Society of Archivists; Australian Library and Information Association; GAIN Australia; DAMA Australia; Health Information Management Association Australia; and RIMPA Global).
Reflecting on the impact of AI on data, information and records across numerous industries, government and private enterprises, the collective voice of government regulators, professional bodies and information management professionals provided some strong and consistent messages to be heeded.
AI is only as trustworthy as the information behind it
AI does not create trust – information quality does. AI cannot compensate for poor-quality, fragmented or unreliable data. AI systems use the data models they are based on (which is why you see the message ‘AI-generated content may be incorrect’ when you use AI models). While AI technologies can be praised for their sophisticated methods, true value is tested by the integrity and veracity of the data they generate. The need for quality data is heightened not diminished by AI.
Information management is not administrative, it’s strategic
An organisation’s approach to AI, it’s risks and effectiveness have become a Board governance function while information and records management is part of an organisation’s strategic infrastructure that underpins service delivery, decision-making, compliance, transparency and responsible innovation. This was reinforced by both the NSW Information and Privacy Commission, and the Qld OIC in their IAW posts this week. Consistent with my presentation at last week’s Australian Healthcare Week, my challenge is for those reading this article to do a stocktake of the extent to which data in your organisation is being treated as a strategic asset. Strategic assets are not forgotten, ignored or dismissed, but are supported by organisational priority, accountability, capacity and capability.
Trust depends on transparency, not technical sophistication
A number of speakers, including Alice Linacre, the Freedom of Information Commissioner reinforced that public trust is shaped by transparency and explainability. The audience was challenged to consider how transparent their organisation is with the public / their consumers on how AI is being used as it relates to processes and functions relating to them. Information governance resulting in clear records, auditable decisions and accessible information is what makes transparency possible in a digital and AI-enabled environment. The OECD Trust Survey conducted in 2024 indicates that while Australia has relatively strong trust foundations, the current information ecosystem has made it harder for individuals to understand and assess the trustworthiness of information, suggesting that strong governance of data use is important to build and maintain trust.
Information governance is essential to ethical AI, not optional
There was a united view that information management and governance principles are continuously applicable regardless of the context. Similarly in healthcare, clinical governance is continuously applicable in AI contexts. A productive and effective median needs to be reached that enables governance compliance while enabling AI solutions to produce effective outcomes.
Information professionals have a key role
With data as a strategic asset, the trustworthiness of data impacting the effectiveness of AI, and the ongoing need for trust, it is clear that information governance and the expert professionals that administer it, have a vital role to play in order for AI outcomes to be effective and purposeful. This is particularly important in the regulatory, clinical safety and clinical quality context of healthcare. Organisations providing and servicing healthcare benefit from the expertise of Health Information Managers to guide information governance including relating to AI. HIMAA’s recent position statement on The Critical Role of Health Information Managers in Digital Health reinforces the value of this important segment of the health workforce.
YES: Although this has not always been the case, the organisation is beginning to understand how important data is for a person centred predictive and preventative future. One that has research as an adjunct tool for the generation of knowledge and evidence, with the health consumer coupled as a citizen scientist.