As Australia marks 50 years of NAIDOC Week, the message aimed at the people building artificial intelligence (AI) is blunt: the technology is arriving, but it has not reckoned with a basic principle often summarized as “nothing about us, without us.” That line matters because AI, like so many technologies before it, can quietly replicate an old pattern. It can take knowledge, convert it into value, and leave the knowledge holders with none of the consent, credit, or return they should have been owed from day one.

This is not a hypothetical ethical debate for the future. It is a problem that emerges at the exact moment AI systems are trained and deployed: when data is gathered, labeled, used, and turned into product decisions. The concern described alongside NAIDOC Week is that AI could become another extractive force. In other words, “taking” can be baked into the pipeline, long before any public statement is made about who benefits. If Indigenous Knowledges are used without consent, without appropriate credit, and without return, the technology can widen a familiar imbalance: the people who supply knowledge become background characters in the story, while organizations that monetize outputs become the main characters.

To understand why this is such a board-level issue, zoom out. AI is built on consumption of information. Whether that information is text, images, or other digital artifacts, AI systems generally learn correlations and patterns from whatever inputs are available. That means incentives are naturally misaligned if organizations treat Indigenous Knowledges as raw material rather than as living, owned, and governed knowledge. Even when teams believe they are “just using data,” they can still create harm if the process ignores Indigenous authority and the principle “nothing about us, without us.” In governance terms, this is the difference between a dataset and a relationship.

Australia’s NAIDOC Week milestone adds weight to the warning because it signals continuity. Honoring “the world’s oldest living culture” is not a ceremonial detail. It is a reminder that Indigenous cultures are not frozen in time and not a source to be harvested once AI goes looking for training material. The article’s framing points to a broader societal reckoning. The last several technological waves brought their own extraction logic. Now AI is being asked, in effect, to do something rare: learn from those past mistakes rather than repeat them.

There is also a second-order risk that executives should not ignore: credibility and trust become operational assets. If AI initiatives are built with a “take-first” mindset, the reputational damage can show up in partnerships, procurement choices, community responses, and long-term adoption. Even if teams believe they are complying with whatever rules exist, the article highlights a principle that goes beyond mere compliance. It is about consent and reciprocity. That is the kind of standard that can become a de facto requirement for the legitimacy of AI projects, especially those touching culture, history, language, and land-linked knowledge.

From a regulatory and policy angle, the key takeaway is the direction of travel: regulators and institutions increasingly want AI development to be accountable to affected communities. While this specific piece does not enumerate particular laws or penalties, it anchors the issue in a widely used governance idea: nothing about marginalized people should be done without their involvement. For decision-makers, this suggests that AI governance frameworks will likely be evaluated not only on technical safety, but also on procedural fairness and community authorization.

So what should executives do with this? First, treat Indigenous Knowledges as governed inputs, not as interchangeable training data. The principle stated alongside NAIDOC Week explicitly calls out what is missing in extractive approaches: consent, credit, and return. Any AI program that cannot answer how it obtains consent, how it attributes appropriately, and how it ensures return will be operating in the risk zone the article warns about.

Second, build “with” rather than “about.” If the headline principle is “nothing about us, without us,” then engagement cannot be a late-stage consultation after decisions are already locked. It has to be integrated into how data is collected, how outputs are evaluated, and how benefits are shared. That is the strategic stakes for boards and leaders of AI-heavy companies: the legitimacy of your models and the sustainability of your AI strategy will increasingly depend on whether you can demonstrate reciprocal collaboration with Indigenous Knowledge holders.

Finally, consider the competitive angle that comes with doing it right. When AI systems are built respectfully, with Indigenous authority and appropriate reciprocity, they can avoid the costly churn of backlash and redesign. When they are built extractively, they can inherit the same structural problems seen in earlier technology cycles. The NAIDOC Week message is essentially a leadership test for the AI era: will this wave of innovation follow the old playbook, or will it adopt consent, credit, and return as first principles?