AI or not, credibility depends on human responsibility - SCMP

Originally published - SCMP on 6 April 2026

I agree with the letter, “Future of journalism should be defined by humans standing behind facts” (March 15), that the enduring value of journalism lies in verification and accountability rather than the mechanical act of writing. That point, however, reflects a broader principle that applies well beyond journalism and is increasingly important in an age of artificial intelligence.

In many areas of public and professional life, legitimacy has never depended on who produced the first draft or what tools were used to generate it. Drafting has always been collaborative and assisted by tools, templates and prior material. What gives a document, decision or statement its authority is the moment it is reviewed, adopted and treated as definitive by someone who bears responsibility for it.

Artificial intelligence does not alter this structure. It simply makes the production of content faster and cheaper. As a result, output alone can no longer serve as a reliable signal of care, understanding or sound judgment. That does not make AI‑assisted content inherently suspect. It means that responsibility must be located elsewhere, precisely where it has always belonged.

What matters is whether someone has checked the content against reliable sources, confirmed its accuracy, understood its implications and is willing to stand behind it if it proves wrong. If that step is taken, the method of creation becomes irrelevant. If it is not taken, the problem is not the technology, but a failure of due diligence and responsibility.

This distinction is important because debates about AI often drift towards questions of authorship or purity of process. Those concerns miss the real issue. Credibility does not rest on who typed the words, how long it took or whether a tool was involved. It rests on adoption, verification and accountability.

Indeed, as the volume of generated content grows, these qualities become more valuable, not less. When information can be produced instantly and distributed widely, the risk of error, distortion and overconfidence increases. Trust, therefore, flows to individuals and institutions that can demonstrate rigorous standards, transparent review processes and clear ownership of outcomes.

Technology may assist the process, but credibility still depends on human responsibility. That principle has not changed, and it is the one worth defending as AI continues to evolve.

Ahmed Ashfaq, Tsim Sha Tsui