From What AI Can Do to How AI Should Be Governed

For much of the past decade, the global conversation about artificial intelligence has been driven by one central question: What can AI do? The answers have been impressive. AI can generate content, analyse vast datasets, detect fraud, assist doctors, automate customer service, screen job applications, accelerate legal research, and even support scientific discovery. Every week seems to bring another breakthrough that pushes the boundaries of what machines can accomplish.

Yet, a noticeable shift is emerging in public discourse. The most important question is no longer about AI’s capabilities but about its governance. Increasingly, policymakers, business leaders, academics, and civil society are asking a different question: How should AI be governed?

A review of recent Facebook discussions reveals that conversations are becoming less fascinated with AI’s technical possibilities and more concerned with its societal implications. Transparency, accountability, human oversight, data governance, security, fairness, and regulation now dominate the discussion. This evolution signals that public understanding of AI is maturing.

One of the strongest themes emerging from these discussions is the recognition that governance cannot be treated as an afterthought. Organisations may find it easy to purchase AI tools, but governing how those tools are used responsibly is proving far more difficult. Contributors repeatedly emphasise that AI governance encompasses data ownership, model validation, bias testing, audit trails, cybersecurity, vendor management, regulatory compliance, and incident response. These are no longer viewed as technical checklists but as strategic imperatives that determine whether AI can be trusted.

Trust has quietly become AI’s most valuable currency. This explains why transparency is receiving unprecedented attention. Proposed policies requiring organisations to disclose when customers are interacting with AI reflect a broader societal expectation that individuals have a right to know when automated systems influence decisions affecting their lives. Whether in banking, healthcare, recruitment, insurance, or media, transparency is increasingly being framed not merely as good corporate practice but as a matter of consumer rights and democratic accountability.

Equally significant is the growing insistence that AI should support, not replace, human judgement. Across sectors, there is widespread agreement that high-stakes decisions must remain subject to meaningful human oversight. AI may assist professionals by identifying patterns, generating recommendations, or improving efficiency, but accountability ultimately rests with people. This reflects a growing consensus that responsible AI is fundamentally human-centred rather than machine-centred.

The recent controversies surrounding AI-generated policy documents further reinforce this lesson. Instances where official documents contained fabricated AI-generated references demonstrate that even institutions responsible for regulating AI are vulnerable when human verification is neglected. Such incidents illustrate that technological sophistication cannot compensate for weak governance. In fact, they highlight a deeper challenge: digital transformation requires institutional competence as much as technological innovation.

Another keynote of current discourse is the expansion of AI risk beyond traditional concerns. Conversations now include cybersecurity vulnerabilities, biological threats, misinformation, loss of human control, and automated AI research capable of accelerating its own development. These discussions suggest that society increasingly views AI through the lens of risk management rather than technological optimism. Innovation remains desirable, but only when accompanied by safeguards capable of preventing unintended consequences.

Governments are also emerging as central actors in this evolving landscape. National AI strategies, parliamentary debates, international summits, and regulatory proposals demonstrate that AI governance is becoming a defining public policy issue. Countries are no longer competing solely to adopt AI technologies; they are also competing to develop governance frameworks that balance innovation with accountability. This represents an important transition from viewing AI as a purely commercial technology to recognising it as critical national infrastructure requiring public oversight.

Perhaps the most encouraging aspect of these conversations is that governance is no longer seen as an obstacle to innovation. Instead, it is increasingly understood as the foundation upon which sustainable innovation depends. Organisations that invest in transparency, ethical safeguards, robust oversight, and responsible deployment are more likely to earn public confidence than those pursuing rapid adoption without adequate controls.

The future of artificial intelligence will not be determined by algorithms alone. It will be shaped by the quality of our institutions, the strength of our governance frameworks, and our willingness to place human values at the centre of technological progress. AI may continue to evolve at extraordinary speed, but public trust will evolve much more slowly—and once lost, it is difficult to regain.

The conversation has changed, and rightly so. The defining question of the AI era is no longer What can AI do? We already know the answer is “almost anything.” The question that will determine whether AI truly benefits society is far more important: How should AI be governed? The countries, organisations, and leaders that answer this question well will not simply lead the next wave of technological innovation; they will shape the future of responsible digital transformation itself.